fix(ask-user): keep question payloads compact (#1208)

Long decision-brief bodies packed into AskUserQuestion's `question` field
and multi-tab batches make the question card unreadable in panel hosts
(VSCode). The brief belongs in markdown before the call; the tool payload
should stay compact.

generate-ask-user-format.ts now documents, in the Format section:
- the two-part contract (markdown brief, then a compact tool_use payload);
- "Tool payload rules": `question` is one sentence / no newlines / <=80
  chars, brief content stays in markdown, batch at most two related
  questions/tabs (sequence the rest), and no duplicated trade-off text
  between `question` and `options[].description`;
- four matching self-check items.

This grafts only the compact-payload guidance onto the current resolver —
it preserves the Conductor / runtime-failure-fallback content that landed
on main since this PR was first opened.

Regenerated all host SKILL.md variants (--host all) and refreshed the
claude/codex/factory ship golden fixtures so the generated docs carry the
new rules.

Tests: new resolver assertions (compact-question, tab-batching,
no-duplicate-tradeoff, self-check items) and a gen-skill-docs assertion
that the rules reach generated tier 2+ SKILL.md. Affected suites
(resolver-ask-user-format, gen-skill-docs, host-config) green.

Fixes #1208

Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>

autoplan/SKILL.md

@@ -345,7 +345,7 @@ Layout: a `D<N>` title + a one-line note to reply with a letter (in Conductor th
 
 ### Format
 
-Every AskUserQuestion is a decision brief and must be sent as tool_use, not prose — unless the documented failure fallback above applies (interactive session + the call is unavailable/erroring), in which case the prose fallback is the correct output.
+Every AskUserQuestion decision has two parts: a markdown decision brief before the call, then a compact tool_use payload. Do not pack the full brief into the tool's `question` string. Every AskUserQuestion is a decision brief and must be sent as tool_use, not prose — unless the documented failure fallback above applies (interactive session + the call is unavailable/erroring), in which case the prose fallback is the correct output.
 
 ```
 D<N> — <one-line question title>
@@ -378,6 +378,12 @@ Effort both-scales: when an option involves effort, label both human-team and CC
 
 Net line closes the tradeoff. Per-skill instructions may add stricter rules.
 
+Tool payload rules:
+- `question` is only the decision prompt: one sentence, no newlines, <=80 chars.
+- Background, regrounding, ELI10, stakes, recommendation, pros/cons, and trade-off tables stay in the markdown brief before the tool call.
+- Ask one decision per tool call when possible; batch at most two related questions/tabs. Sequence independent decisions instead of sending 3+ tabs.
+- Do not duplicate the same trade-off text in both `question` and `options[].description`. Prefer putting option-specific trade-offs in `options[].description`.
+
 ### Handling 5+ options — split, never drop
 
 AskUserQuestion caps every call at **4 options**. With 5+ real options, NEVER
@@ -425,6 +431,10 @@ Before calling AskUserQuestion, verify:
 - [ ] (recommended) label on one option (even for neutral-posture)
 - [ ] Dual-scale effort labels on effort-bearing options (human / CC)
 - [ ] Net line closes the decision
+- [ ] `question` is one sentence, no newlines, <=80 chars
+- [ ] Tool call has no more than two related questions/tabs
+- [ ] No duplicated trade-off text between `question` and `options[].description`
+- [ ] You wrote the brief, then called the tool_use payload
 - [ ] You are calling the tool, not writing prose — unless `CONDUCTOR_SESSION: true` (then prose is the DEFAULT, not the tool) OR the documented failure fallback applies (then: prose with the mandatory triad — issue ELI10, per-choice Completeness, Recommendation + `(recommended)` — and a "reply with a letter" instruction, then STOP)
 - [ ] Non-ASCII characters (CJK / accents) written directly, NOT \u-escaped
 - [ ] If you had 5+ options, you split (or batched into ≤4-groups) — did NOT drop any

canary/SKILL.md

@@ -337,7 +337,7 @@ Layout: a `D<N>` title + a one-line note to reply with a letter (in Conductor th
 
 ### Format
 
-Every AskUserQuestion is a decision brief and must be sent as tool_use, not prose — unless the documented failure fallback above applies (interactive session + the call is unavailable/erroring), in which case the prose fallback is the correct output.
+Every AskUserQuestion decision has two parts: a markdown decision brief before the call, then a compact tool_use payload. Do not pack the full brief into the tool's `question` string. Every AskUserQuestion is a decision brief and must be sent as tool_use, not prose — unless the documented failure fallback above applies (interactive session + the call is unavailable/erroring), in which case the prose fallback is the correct output.
 
 ```
 D<N> — <one-line question title>
@@ -370,6 +370,12 @@ Effort both-scales: when an option involves effort, label both human-team and CC
 
 Net line closes the tradeoff. Per-skill instructions may add stricter rules.
 
+Tool payload rules:
+- `question` is only the decision prompt: one sentence, no newlines, <=80 chars.
+- Background, regrounding, ELI10, stakes, recommendation, pros/cons, and trade-off tables stay in the markdown brief before the tool call.
+- Ask one decision per tool call when possible; batch at most two related questions/tabs. Sequence independent decisions instead of sending 3+ tabs.
+- Do not duplicate the same trade-off text in both `question` and `options[].description`. Prefer putting option-specific trade-offs in `options[].description`.
+
 ### Handling 5+ options — split, never drop
 
 AskUserQuestion caps every call at **4 options**. With 5+ real options, NEVER
@@ -417,6 +423,10 @@ Before calling AskUserQuestion, verify:
 - [ ] (recommended) label on one option (even for neutral-posture)
 - [ ] Dual-scale effort labels on effort-bearing options (human / CC)
 - [ ] Net line closes the decision
+- [ ] `question` is one sentence, no newlines, <=80 chars
+- [ ] Tool call has no more than two related questions/tabs
+- [ ] No duplicated trade-off text between `question` and `options[].description`
+- [ ] You wrote the brief, then called the tool_use payload
 - [ ] You are calling the tool, not writing prose — unless `CONDUCTOR_SESSION: true` (then prose is the DEFAULT, not the tool) OR the documented failure fallback applies (then: prose with the mandatory triad — issue ELI10, per-choice Completeness, Recommendation + `(recommended)` — and a "reply with a letter" instruction, then STOP)
 - [ ] Non-ASCII characters (CJK / accents) written directly, NOT \u-escaped
 - [ ] If you had 5+ options, you split (or batched into ≤4-groups) — did NOT drop any

codex/SKILL.md

@@ -340,7 +340,7 @@ Layout: a `D<N>` title + a one-line note to reply with a letter (in Conductor th
 
 ### Format
 
-Every AskUserQuestion is a decision brief and must be sent as tool_use, not prose — unless the documented failure fallback above applies (interactive session + the call is unavailable/erroring), in which case the prose fallback is the correct output.
+Every AskUserQuestion decision has two parts: a markdown decision brief before the call, then a compact tool_use payload. Do not pack the full brief into the tool's `question` string. Every AskUserQuestion is a decision brief and must be sent as tool_use, not prose — unless the documented failure fallback above applies (interactive session + the call is unavailable/erroring), in which case the prose fallback is the correct output.
 
 ```
 D<N> — <one-line question title>
@@ -373,6 +373,12 @@ Effort both-scales: when an option involves effort, label both human-team and CC
 
 Net line closes the tradeoff. Per-skill instructions may add stricter rules.
 
+Tool payload rules:
+- `question` is only the decision prompt: one sentence, no newlines, <=80 chars.
+- Background, regrounding, ELI10, stakes, recommendation, pros/cons, and trade-off tables stay in the markdown brief before the tool call.
+- Ask one decision per tool call when possible; batch at most two related questions/tabs. Sequence independent decisions instead of sending 3+ tabs.
+- Do not duplicate the same trade-off text in both `question` and `options[].description`. Prefer putting option-specific trade-offs in `options[].description`.
+
 ### Handling 5+ options — split, never drop
 
 AskUserQuestion caps every call at **4 options**. With 5+ real options, NEVER
@@ -420,6 +426,10 @@ Before calling AskUserQuestion, verify:
 - [ ] (recommended) label on one option (even for neutral-posture)
 - [ ] Dual-scale effort labels on effort-bearing options (human / CC)
 - [ ] Net line closes the decision
+- [ ] `question` is one sentence, no newlines, <=80 chars
+- [ ] Tool call has no more than two related questions/tabs
+- [ ] No duplicated trade-off text between `question` and `options[].description`
+- [ ] You wrote the brief, then called the tool_use payload
 - [ ] You are calling the tool, not writing prose — unless `CONDUCTOR_SESSION: true` (then prose is the DEFAULT, not the tool) OR the documented failure fallback applies (then: prose with the mandatory triad — issue ELI10, per-choice Completeness, Recommendation + `(recommended)` — and a "reply with a letter" instruction, then STOP)
 - [ ] Non-ASCII characters (CJK / accents) written directly, NOT \u-escaped
 - [ ] If you had 5+ options, you split (or batched into ≤4-groups) — did NOT drop any

context-restore/SKILL.md

@@ -341,7 +341,7 @@ Layout: a `D<N>` title + a one-line note to reply with a letter (in Conductor th
 
 ### Format
 
-Every AskUserQuestion is a decision brief and must be sent as tool_use, not prose — unless the documented failure fallback above applies (interactive session + the call is unavailable/erroring), in which case the prose fallback is the correct output.
+Every AskUserQuestion decision has two parts: a markdown decision brief before the call, then a compact tool_use payload. Do not pack the full brief into the tool's `question` string. Every AskUserQuestion is a decision brief and must be sent as tool_use, not prose — unless the documented failure fallback above applies (interactive session + the call is unavailable/erroring), in which case the prose fallback is the correct output.
 
 ```
 D<N> — <one-line question title>
@@ -374,6 +374,12 @@ Effort both-scales: when an option involves effort, label both human-team and CC
 
 Net line closes the tradeoff. Per-skill instructions may add stricter rules.
 
+Tool payload rules:
+- `question` is only the decision prompt: one sentence, no newlines, <=80 chars.
+- Background, regrounding, ELI10, stakes, recommendation, pros/cons, and trade-off tables stay in the markdown brief before the tool call.
+- Ask one decision per tool call when possible; batch at most two related questions/tabs. Sequence independent decisions instead of sending 3+ tabs.
+- Do not duplicate the same trade-off text in both `question` and `options[].description`. Prefer putting option-specific trade-offs in `options[].description`.
+
 ### Handling 5+ options — split, never drop
 
 AskUserQuestion caps every call at **4 options**. With 5+ real options, NEVER
@@ -421,6 +427,10 @@ Before calling AskUserQuestion, verify:
 - [ ] (recommended) label on one option (even for neutral-posture)
 - [ ] Dual-scale effort labels on effort-bearing options (human / CC)
 - [ ] Net line closes the decision
+- [ ] `question` is one sentence, no newlines, <=80 chars
+- [ ] Tool call has no more than two related questions/tabs
+- [ ] No duplicated trade-off text between `question` and `options[].description`
+- [ ] You wrote the brief, then called the tool_use payload
 - [ ] You are calling the tool, not writing prose — unless `CONDUCTOR_SESSION: true` (then prose is the DEFAULT, not the tool) OR the documented failure fallback applies (then: prose with the mandatory triad — issue ELI10, per-choice Completeness, Recommendation + `(recommended)` — and a "reply with a letter" instruction, then STOP)
 - [ ] Non-ASCII characters (CJK / accents) written directly, NOT \u-escaped
 - [ ] If you had 5+ options, you split (or batched into ≤4-groups) — did NOT drop any

context-save/SKILL.md

@@ -340,7 +340,7 @@ Layout: a `D<N>` title + a one-line note to reply with a letter (in Conductor th
 
 ### Format
 
-Every AskUserQuestion is a decision brief and must be sent as tool_use, not prose — unless the documented failure fallback above applies (interactive session + the call is unavailable/erroring), in which case the prose fallback is the correct output.
+Every AskUserQuestion decision has two parts: a markdown decision brief before the call, then a compact tool_use payload. Do not pack the full brief into the tool's `question` string. Every AskUserQuestion is a decision brief and must be sent as tool_use, not prose — unless the documented failure fallback above applies (interactive session + the call is unavailable/erroring), in which case the prose fallback is the correct output.
 
 ```
 D<N> — <one-line question title>
@@ -373,6 +373,12 @@ Effort both-scales: when an option involves effort, label both human-team and CC
 
 Net line closes the tradeoff. Per-skill instructions may add stricter rules.
 
+Tool payload rules:
+- `question` is only the decision prompt: one sentence, no newlines, <=80 chars.
+- Background, regrounding, ELI10, stakes, recommendation, pros/cons, and trade-off tables stay in the markdown brief before the tool call.
+- Ask one decision per tool call when possible; batch at most two related questions/tabs. Sequence independent decisions instead of sending 3+ tabs.
+- Do not duplicate the same trade-off text in both `question` and `options[].description`. Prefer putting option-specific trade-offs in `options[].description`.
+
 ### Handling 5+ options — split, never drop
 
 AskUserQuestion caps every call at **4 options**. With 5+ real options, NEVER
@@ -420,6 +426,10 @@ Before calling AskUserQuestion, verify:
 - [ ] (recommended) label on one option (even for neutral-posture)
 - [ ] Dual-scale effort labels on effort-bearing options (human / CC)
 - [ ] Net line closes the decision
+- [ ] `question` is one sentence, no newlines, <=80 chars
+- [ ] Tool call has no more than two related questions/tabs
+- [ ] No duplicated trade-off text between `question` and `options[].description`
+- [ ] You wrote the brief, then called the tool_use payload
 - [ ] You are calling the tool, not writing prose — unless `CONDUCTOR_SESSION: true` (then prose is the DEFAULT, not the tool) OR the documented failure fallback applies (then: prose with the mandatory triad — issue ELI10, per-choice Completeness, Recommendation + `(recommended)` — and a "reply with a letter" instruction, then STOP)
 - [ ] Non-ASCII characters (CJK / accents) written directly, NOT \u-escaped
 - [ ] If you had 5+ options, you split (or batched into ≤4-groups) — did NOT drop any

cso/SKILL.md

@@ -343,7 +343,7 @@ Layout: a `D<N>` title + a one-line note to reply with a letter (in Conductor th
 
 ### Format
 
-Every AskUserQuestion is a decision brief and must be sent as tool_use, not prose — unless the documented failure fallback above applies (interactive session + the call is unavailable/erroring), in which case the prose fallback is the correct output.
+Every AskUserQuestion decision has two parts: a markdown decision brief before the call, then a compact tool_use payload. Do not pack the full brief into the tool's `question` string. Every AskUserQuestion is a decision brief and must be sent as tool_use, not prose — unless the documented failure fallback above applies (interactive session + the call is unavailable/erroring), in which case the prose fallback is the correct output.
 
 ```
 D<N> — <one-line question title>
@@ -376,6 +376,12 @@ Effort both-scales: when an option involves effort, label both human-team and CC
 
 Net line closes the tradeoff. Per-skill instructions may add stricter rules.
 
+Tool payload rules:
+- `question` is only the decision prompt: one sentence, no newlines, <=80 chars.
+- Background, regrounding, ELI10, stakes, recommendation, pros/cons, and trade-off tables stay in the markdown brief before the tool call.
+- Ask one decision per tool call when possible; batch at most two related questions/tabs. Sequence independent decisions instead of sending 3+ tabs.
+- Do not duplicate the same trade-off text in both `question` and `options[].description`. Prefer putting option-specific trade-offs in `options[].description`.
+
 ### Handling 5+ options — split, never drop
 
 AskUserQuestion caps every call at **4 options**. With 5+ real options, NEVER
@@ -423,6 +429,10 @@ Before calling AskUserQuestion, verify:
 - [ ] (recommended) label on one option (even for neutral-posture)
 - [ ] Dual-scale effort labels on effort-bearing options (human / CC)
 - [ ] Net line closes the decision
+- [ ] `question` is one sentence, no newlines, <=80 chars
+- [ ] Tool call has no more than two related questions/tabs
+- [ ] No duplicated trade-off text between `question` and `options[].description`
+- [ ] You wrote the brief, then called the tool_use payload
 - [ ] You are calling the tool, not writing prose — unless `CONDUCTOR_SESSION: true` (then prose is the DEFAULT, not the tool) OR the documented failure fallback applies (then: prose with the mandatory triad — issue ELI10, per-choice Completeness, Recommendation + `(recommended)` — and a "reply with a letter" instruction, then STOP)
 - [ ] Non-ASCII characters (CJK / accents) written directly, NOT \u-escaped
 - [ ] If you had 5+ options, you split (or batched into ≤4-groups) — did NOT drop any

design-consultation/SKILL.md

@@ -363,7 +363,7 @@ Layout: a `D<N>` title + a one-line note to reply with a letter (in Conductor th
 
 ### Format
 
-Every AskUserQuestion is a decision brief and must be sent as tool_use, not prose — unless the documented failure fallback above applies (interactive session + the call is unavailable/erroring), in which case the prose fallback is the correct output.
+Every AskUserQuestion decision has two parts: a markdown decision brief before the call, then a compact tool_use payload. Do not pack the full brief into the tool's `question` string. Every AskUserQuestion is a decision brief and must be sent as tool_use, not prose — unless the documented failure fallback above applies (interactive session + the call is unavailable/erroring), in which case the prose fallback is the correct output.
 
 ```
 D<N> — <one-line question title>
@@ -396,6 +396,12 @@ Effort both-scales: when an option involves effort, label both human-team and CC
 
 Net line closes the tradeoff. Per-skill instructions may add stricter rules.
 
+Tool payload rules:
+- `question` is only the decision prompt: one sentence, no newlines, <=80 chars.
+- Background, regrounding, ELI10, stakes, recommendation, pros/cons, and trade-off tables stay in the markdown brief before the tool call.
+- Ask one decision per tool call when possible; batch at most two related questions/tabs. Sequence independent decisions instead of sending 3+ tabs.
+- Do not duplicate the same trade-off text in both `question` and `options[].description`. Prefer putting option-specific trade-offs in `options[].description`.
+
 ### Handling 5+ options — split, never drop
 
 AskUserQuestion caps every call at **4 options**. With 5+ real options, NEVER
@@ -443,6 +449,10 @@ Before calling AskUserQuestion, verify:
 - [ ] (recommended) label on one option (even for neutral-posture)
 - [ ] Dual-scale effort labels on effort-bearing options (human / CC)
 - [ ] Net line closes the decision
+- [ ] `question` is one sentence, no newlines, <=80 chars
+- [ ] Tool call has no more than two related questions/tabs
+- [ ] No duplicated trade-off text between `question` and `options[].description`
+- [ ] You wrote the brief, then called the tool_use payload
 - [ ] You are calling the tool, not writing prose — unless `CONDUCTOR_SESSION: true` (then prose is the DEFAULT, not the tool) OR the documented failure fallback applies (then: prose with the mandatory triad — issue ELI10, per-choice Completeness, Recommendation + `(recommended)` — and a "reply with a letter" instruction, then STOP)
 - [ ] Non-ASCII characters (CJK / accents) written directly, NOT \u-escaped
 - [ ] If you had 5+ options, you split (or batched into ≤4-groups) — did NOT drop any

design-html/SKILL.md

@@ -344,7 +344,7 @@ Layout: a `D<N>` title + a one-line note to reply with a letter (in Conductor th
 
 ### Format
 
-Every AskUserQuestion is a decision brief and must be sent as tool_use, not prose — unless the documented failure fallback above applies (interactive session + the call is unavailable/erroring), in which case the prose fallback is the correct output.
+Every AskUserQuestion decision has two parts: a markdown decision brief before the call, then a compact tool_use payload. Do not pack the full brief into the tool's `question` string. Every AskUserQuestion is a decision brief and must be sent as tool_use, not prose — unless the documented failure fallback above applies (interactive session + the call is unavailable/erroring), in which case the prose fallback is the correct output.
 
 ```
 D<N> — <one-line question title>
@@ -377,6 +377,12 @@ Effort both-scales: when an option involves effort, label both human-team and CC
 
 Net line closes the tradeoff. Per-skill instructions may add stricter rules.
 
+Tool payload rules:
+- `question` is only the decision prompt: one sentence, no newlines, <=80 chars.
+- Background, regrounding, ELI10, stakes, recommendation, pros/cons, and trade-off tables stay in the markdown brief before the tool call.
+- Ask one decision per tool call when possible; batch at most two related questions/tabs. Sequence independent decisions instead of sending 3+ tabs.
+- Do not duplicate the same trade-off text in both `question` and `options[].description`. Prefer putting option-specific trade-offs in `options[].description`.
+
 ### Handling 5+ options — split, never drop
 
 AskUserQuestion caps every call at **4 options**. With 5+ real options, NEVER
@@ -424,6 +430,10 @@ Before calling AskUserQuestion, verify:
 - [ ] (recommended) label on one option (even for neutral-posture)
 - [ ] Dual-scale effort labels on effort-bearing options (human / CC)
 - [ ] Net line closes the decision
+- [ ] `question` is one sentence, no newlines, <=80 chars
+- [ ] Tool call has no more than two related questions/tabs
+- [ ] No duplicated trade-off text between `question` and `options[].description`
+- [ ] You wrote the brief, then called the tool_use payload
 - [ ] You are calling the tool, not writing prose — unless `CONDUCTOR_SESSION: true` (then prose is the DEFAULT, not the tool) OR the documented failure fallback applies (then: prose with the mandatory triad — issue ELI10, per-choice Completeness, Recommendation + `(recommended)` — and a "reply with a letter" instruction, then STOP)
 - [ ] Non-ASCII characters (CJK / accents) written directly, NOT \u-escaped
 - [ ] If you had 5+ options, you split (or batched into ≤4-groups) — did NOT drop any

design-review/SKILL.md

@@ -341,7 +341,7 @@ Layout: a `D<N>` title + a one-line note to reply with a letter (in Conductor th
 
 ### Format
 
-Every AskUserQuestion is a decision brief and must be sent as tool_use, not prose — unless the documented failure fallback above applies (interactive session + the call is unavailable/erroring), in which case the prose fallback is the correct output.
+Every AskUserQuestion decision has two parts: a markdown decision brief before the call, then a compact tool_use payload. Do not pack the full brief into the tool's `question` string. Every AskUserQuestion is a decision brief and must be sent as tool_use, not prose — unless the documented failure fallback above applies (interactive session + the call is unavailable/erroring), in which case the prose fallback is the correct output.
 
 ```
 D<N> — <one-line question title>
@@ -374,6 +374,12 @@ Effort both-scales: when an option involves effort, label both human-team and CC
 
 Net line closes the tradeoff. Per-skill instructions may add stricter rules.
 
+Tool payload rules:
+- `question` is only the decision prompt: one sentence, no newlines, <=80 chars.
+- Background, regrounding, ELI10, stakes, recommendation, pros/cons, and trade-off tables stay in the markdown brief before the tool call.
+- Ask one decision per tool call when possible; batch at most two related questions/tabs. Sequence independent decisions instead of sending 3+ tabs.
+- Do not duplicate the same trade-off text in both `question` and `options[].description`. Prefer putting option-specific trade-offs in `options[].description`.
+
 ### Handling 5+ options — split, never drop
 
 AskUserQuestion caps every call at **4 options**. With 5+ real options, NEVER
@@ -421,6 +427,10 @@ Before calling AskUserQuestion, verify:
 - [ ] (recommended) label on one option (even for neutral-posture)
 - [ ] Dual-scale effort labels on effort-bearing options (human / CC)
 - [ ] Net line closes the decision
+- [ ] `question` is one sentence, no newlines, <=80 chars
+- [ ] Tool call has no more than two related questions/tabs
+- [ ] No duplicated trade-off text between `question` and `options[].description`
+- [ ] You wrote the brief, then called the tool_use payload
 - [ ] You are calling the tool, not writing prose — unless `CONDUCTOR_SESSION: true` (then prose is the DEFAULT, not the tool) OR the documented failure fallback applies (then: prose with the mandatory triad — issue ELI10, per-choice Completeness, Recommendation + `(recommended)` — and a "reply with a letter" instruction, then STOP)
 - [ ] Non-ASCII characters (CJK / accents) written directly, NOT \u-escaped
 - [ ] If you had 5+ options, you split (or batched into ≤4-groups) — did NOT drop any

design-shotgun/SKILL.md

@@ -358,7 +358,7 @@ Layout: a `D<N>` title + a one-line note to reply with a letter (in Conductor th
 
 ### Format
 
-Every AskUserQuestion is a decision brief and must be sent as tool_use, not prose — unless the documented failure fallback above applies (interactive session + the call is unavailable/erroring), in which case the prose fallback is the correct output.
+Every AskUserQuestion decision has two parts: a markdown decision brief before the call, then a compact tool_use payload. Do not pack the full brief into the tool's `question` string. Every AskUserQuestion is a decision brief and must be sent as tool_use, not prose — unless the documented failure fallback above applies (interactive session + the call is unavailable/erroring), in which case the prose fallback is the correct output.
 
 ```
 D<N> — <one-line question title>
@@ -391,6 +391,12 @@ Effort both-scales: when an option involves effort, label both human-team and CC
 
 Net line closes the tradeoff. Per-skill instructions may add stricter rules.
 
+Tool payload rules:
+- `question` is only the decision prompt: one sentence, no newlines, <=80 chars.
+- Background, regrounding, ELI10, stakes, recommendation, pros/cons, and trade-off tables stay in the markdown brief before the tool call.
+- Ask one decision per tool call when possible; batch at most two related questions/tabs. Sequence independent decisions instead of sending 3+ tabs.
+- Do not duplicate the same trade-off text in both `question` and `options[].description`. Prefer putting option-specific trade-offs in `options[].description`.
+
 ### Handling 5+ options — split, never drop
 
 AskUserQuestion caps every call at **4 options**. With 5+ real options, NEVER
@@ -438,6 +444,10 @@ Before calling AskUserQuestion, verify:
 - [ ] (recommended) label on one option (even for neutral-posture)
 - [ ] Dual-scale effort labels on effort-bearing options (human / CC)
 - [ ] Net line closes the decision
+- [ ] `question` is one sentence, no newlines, <=80 chars
+- [ ] Tool call has no more than two related questions/tabs
+- [ ] No duplicated trade-off text between `question` and `options[].description`
+- [ ] You wrote the brief, then called the tool_use payload
 - [ ] You are calling the tool, not writing prose — unless `CONDUCTOR_SESSION: true` (then prose is the DEFAULT, not the tool) OR the documented failure fallback applies (then: prose with the mandatory triad — issue ELI10, per-choice Completeness, Recommendation + `(recommended)` — and a "reply with a letter" instruction, then STOP)
 - [ ] Non-ASCII characters (CJK / accents) written directly, NOT \u-escaped
 - [ ] If you had 5+ options, you split (or batched into ≤4-groups) — did NOT drop any

devex-review/SKILL.md

@@ -343,7 +343,7 @@ Layout: a `D<N>` title + a one-line note to reply with a letter (in Conductor th
 
 ### Format
 
-Every AskUserQuestion is a decision brief and must be sent as tool_use, not prose — unless the documented failure fallback above applies (interactive session + the call is unavailable/erroring), in which case the prose fallback is the correct output.
+Every AskUserQuestion decision has two parts: a markdown decision brief before the call, then a compact tool_use payload. Do not pack the full brief into the tool's `question` string. Every AskUserQuestion is a decision brief and must be sent as tool_use, not prose — unless the documented failure fallback above applies (interactive session + the call is unavailable/erroring), in which case the prose fallback is the correct output.
 
 ```
 D<N> — <one-line question title>
@@ -376,6 +376,12 @@ Effort both-scales: when an option involves effort, label both human-team and CC
 
 Net line closes the tradeoff. Per-skill instructions may add stricter rules.
 
+Tool payload rules:
+- `question` is only the decision prompt: one sentence, no newlines, <=80 chars.
+- Background, regrounding, ELI10, stakes, recommendation, pros/cons, and trade-off tables stay in the markdown brief before the tool call.
+- Ask one decision per tool call when possible; batch at most two related questions/tabs. Sequence independent decisions instead of sending 3+ tabs.
+- Do not duplicate the same trade-off text in both `question` and `options[].description`. Prefer putting option-specific trade-offs in `options[].description`.
+
 ### Handling 5+ options — split, never drop
 
 AskUserQuestion caps every call at **4 options**. With 5+ real options, NEVER
@@ -423,6 +429,10 @@ Before calling AskUserQuestion, verify:
 - [ ] (recommended) label on one option (even for neutral-posture)
 - [ ] Dual-scale effort labels on effort-bearing options (human / CC)
 - [ ] Net line closes the decision
+- [ ] `question` is one sentence, no newlines, <=80 chars
+- [ ] Tool call has no more than two related questions/tabs
+- [ ] No duplicated trade-off text between `question` and `options[].description`
+- [ ] You wrote the brief, then called the tool_use payload
 - [ ] You are calling the tool, not writing prose — unless `CONDUCTOR_SESSION: true` (then prose is the DEFAULT, not the tool) OR the documented failure fallback applies (then: prose with the mandatory triad — issue ELI10, per-choice Completeness, Recommendation + `(recommended)` — and a "reply with a letter" instruction, then STOP)
 - [ ] Non-ASCII characters (CJK / accents) written directly, NOT \u-escaped
 - [ ] If you had 5+ options, you split (or batched into ≤4-groups) — did NOT drop any

diagram/SKILL.md

@@ -338,7 +338,7 @@ Layout: a `D<N>` title + a one-line note to reply with a letter (in Conductor th
 
 ### Format
 
-Every AskUserQuestion is a decision brief and must be sent as tool_use, not prose — unless the documented failure fallback above applies (interactive session + the call is unavailable/erroring), in which case the prose fallback is the correct output.
+Every AskUserQuestion decision has two parts: a markdown decision brief before the call, then a compact tool_use payload. Do not pack the full brief into the tool's `question` string. Every AskUserQuestion is a decision brief and must be sent as tool_use, not prose — unless the documented failure fallback above applies (interactive session + the call is unavailable/erroring), in which case the prose fallback is the correct output.
 
 ```
 D<N> — <one-line question title>
@@ -371,6 +371,12 @@ Effort both-scales: when an option involves effort, label both human-team and CC
 
 Net line closes the tradeoff. Per-skill instructions may add stricter rules.
 
+Tool payload rules:
+- `question` is only the decision prompt: one sentence, no newlines, <=80 chars.
+- Background, regrounding, ELI10, stakes, recommendation, pros/cons, and trade-off tables stay in the markdown brief before the tool call.
+- Ask one decision per tool call when possible; batch at most two related questions/tabs. Sequence independent decisions instead of sending 3+ tabs.
+- Do not duplicate the same trade-off text in both `question` and `options[].description`. Prefer putting option-specific trade-offs in `options[].description`.
+
 ### Handling 5+ options — split, never drop
 
 AskUserQuestion caps every call at **4 options**. With 5+ real options, NEVER
@@ -418,6 +424,10 @@ Before calling AskUserQuestion, verify:
 - [ ] (recommended) label on one option (even for neutral-posture)
 - [ ] Dual-scale effort labels on effort-bearing options (human / CC)
 - [ ] Net line closes the decision
+- [ ] `question` is one sentence, no newlines, <=80 chars
+- [ ] Tool call has no more than two related questions/tabs
+- [ ] No duplicated trade-off text between `question` and `options[].description`
+- [ ] You wrote the brief, then called the tool_use payload
 - [ ] You are calling the tool, not writing prose — unless `CONDUCTOR_SESSION: true` (then prose is the DEFAULT, not the tool) OR the documented failure fallback applies (then: prose with the mandatory triad — issue ELI10, per-choice Completeness, Recommendation + `(recommended)` — and a "reply with a letter" instruction, then STOP)
 - [ ] Non-ASCII characters (CJK / accents) written directly, NOT \u-escaped
 - [ ] If you had 5+ options, you split (or batched into ≤4-groups) — did NOT drop any

document-generate/SKILL.md

@@ -343,7 +343,7 @@ Layout: a `D<N>` title + a one-line note to reply with a letter (in Conductor th
 
 ### Format
 
-Every AskUserQuestion is a decision brief and must be sent as tool_use, not prose — unless the documented failure fallback above applies (interactive session + the call is unavailable/erroring), in which case the prose fallback is the correct output.
+Every AskUserQuestion decision has two parts: a markdown decision brief before the call, then a compact tool_use payload. Do not pack the full brief into the tool's `question` string. Every AskUserQuestion is a decision brief and must be sent as tool_use, not prose — unless the documented failure fallback above applies (interactive session + the call is unavailable/erroring), in which case the prose fallback is the correct output.
 
 ```
 D<N> — <one-line question title>
@@ -376,6 +376,12 @@ Effort both-scales: when an option involves effort, label both human-team and CC
 
 Net line closes the tradeoff. Per-skill instructions may add stricter rules.
 
+Tool payload rules:
+- `question` is only the decision prompt: one sentence, no newlines, <=80 chars.
+- Background, regrounding, ELI10, stakes, recommendation, pros/cons, and trade-off tables stay in the markdown brief before the tool call.
+- Ask one decision per tool call when possible; batch at most two related questions/tabs. Sequence independent decisions instead of sending 3+ tabs.
+- Do not duplicate the same trade-off text in both `question` and `options[].description`. Prefer putting option-specific trade-offs in `options[].description`.
+
 ### Handling 5+ options — split, never drop
 
 AskUserQuestion caps every call at **4 options**. With 5+ real options, NEVER
@@ -423,6 +429,10 @@ Before calling AskUserQuestion, verify:
 - [ ] (recommended) label on one option (even for neutral-posture)
 - [ ] Dual-scale effort labels on effort-bearing options (human / CC)
 - [ ] Net line closes the decision
+- [ ] `question` is one sentence, no newlines, <=80 chars
+- [ ] Tool call has no more than two related questions/tabs
+- [ ] No duplicated trade-off text between `question` and `options[].description`
+- [ ] You wrote the brief, then called the tool_use payload
 - [ ] You are calling the tool, not writing prose — unless `CONDUCTOR_SESSION: true` (then prose is the DEFAULT, not the tool) OR the documented failure fallback applies (then: prose with the mandatory triad — issue ELI10, per-choice Completeness, Recommendation + `(recommended)` — and a "reply with a letter" instruction, then STOP)
 - [ ] Non-ASCII characters (CJK / accents) written directly, NOT \u-escaped
 - [ ] If you had 5+ options, you split (or batched into ≤4-groups) — did NOT drop any

document-release/SKILL.md

@@ -341,7 +341,7 @@ Layout: a `D<N>` title + a one-line note to reply with a letter (in Conductor th
 
 ### Format
 
-Every AskUserQuestion is a decision brief and must be sent as tool_use, not prose — unless the documented failure fallback above applies (interactive session + the call is unavailable/erroring), in which case the prose fallback is the correct output.
+Every AskUserQuestion decision has two parts: a markdown decision brief before the call, then a compact tool_use payload. Do not pack the full brief into the tool's `question` string. Every AskUserQuestion is a decision brief and must be sent as tool_use, not prose — unless the documented failure fallback above applies (interactive session + the call is unavailable/erroring), in which case the prose fallback is the correct output.
 
 ```
 D<N> — <one-line question title>
@@ -374,6 +374,12 @@ Effort both-scales: when an option involves effort, label both human-team and CC
 
 Net line closes the tradeoff. Per-skill instructions may add stricter rules.
 
+Tool payload rules:
+- `question` is only the decision prompt: one sentence, no newlines, <=80 chars.
+- Background, regrounding, ELI10, stakes, recommendation, pros/cons, and trade-off tables stay in the markdown brief before the tool call.
+- Ask one decision per tool call when possible; batch at most two related questions/tabs. Sequence independent decisions instead of sending 3+ tabs.
+- Do not duplicate the same trade-off text in both `question` and `options[].description`. Prefer putting option-specific trade-offs in `options[].description`.
+
 ### Handling 5+ options — split, never drop
 
 AskUserQuestion caps every call at **4 options**. With 5+ real options, NEVER
@@ -421,6 +427,10 @@ Before calling AskUserQuestion, verify:
 - [ ] (recommended) label on one option (even for neutral-posture)
 - [ ] Dual-scale effort labels on effort-bearing options (human / CC)
 - [ ] Net line closes the decision
+- [ ] `question` is one sentence, no newlines, <=80 chars
+- [ ] Tool call has no more than two related questions/tabs
+- [ ] No duplicated trade-off text between `question` and `options[].description`
+- [ ] You wrote the brief, then called the tool_use payload
 - [ ] You are calling the tool, not writing prose — unless `CONDUCTOR_SESSION: true` (then prose is the DEFAULT, not the tool) OR the documented failure fallback applies (then: prose with the mandatory triad — issue ELI10, per-choice Completeness, Recommendation + `(recommended)` — and a "reply with a letter" instruction, then STOP)
 - [ ] Non-ASCII characters (CJK / accents) written directly, NOT \u-escaped
 - [ ] If you had 5+ options, you split (or batched into ≤4-groups) — did NOT drop any

health/SKILL.md

@@ -339,7 +339,7 @@ Layout: a `D<N>` title + a one-line note to reply with a letter (in Conductor th
 
 ### Format
 
-Every AskUserQuestion is a decision brief and must be sent as tool_use, not prose — unless the documented failure fallback above applies (interactive session + the call is unavailable/erroring), in which case the prose fallback is the correct output.
+Every AskUserQuestion decision has two parts: a markdown decision brief before the call, then a compact tool_use payload. Do not pack the full brief into the tool's `question` string. Every AskUserQuestion is a decision brief and must be sent as tool_use, not prose — unless the documented failure fallback above applies (interactive session + the call is unavailable/erroring), in which case the prose fallback is the correct output.
 
 ```
 D<N> — <one-line question title>
@@ -372,6 +372,12 @@ Effort both-scales: when an option involves effort, label both human-team and CC
 
 Net line closes the tradeoff. Per-skill instructions may add stricter rules.
 
+Tool payload rules:
+- `question` is only the decision prompt: one sentence, no newlines, <=80 chars.
+- Background, regrounding, ELI10, stakes, recommendation, pros/cons, and trade-off tables stay in the markdown brief before the tool call.
+- Ask one decision per tool call when possible; batch at most two related questions/tabs. Sequence independent decisions instead of sending 3+ tabs.
+- Do not duplicate the same trade-off text in both `question` and `options[].description`. Prefer putting option-specific trade-offs in `options[].description`.
+
 ### Handling 5+ options — split, never drop
 
 AskUserQuestion caps every call at **4 options**. With 5+ real options, NEVER
@@ -419,6 +425,10 @@ Before calling AskUserQuestion, verify:
 - [ ] (recommended) label on one option (even for neutral-posture)
 - [ ] Dual-scale effort labels on effort-bearing options (human / CC)
 - [ ] Net line closes the decision
+- [ ] `question` is one sentence, no newlines, <=80 chars
+- [ ] Tool call has no more than two related questions/tabs
+- [ ] No duplicated trade-off text between `question` and `options[].description`
+- [ ] You wrote the brief, then called the tool_use payload
 - [ ] You are calling the tool, not writing prose — unless `CONDUCTOR_SESSION: true` (then prose is the DEFAULT, not the tool) OR the documented failure fallback applies (then: prose with the mandatory triad — issue ELI10, per-choice Completeness, Recommendation + `(recommended)` — and a "reply with a letter" instruction, then STOP)
 - [ ] Non-ASCII characters (CJK / accents) written directly, NOT \u-escaped
 - [ ] If you had 5+ options, you split (or batched into ≤4-groups) — did NOT drop any

investigate/SKILL.md

@@ -378,7 +378,7 @@ Layout: a `D<N>` title + a one-line note to reply with a letter (in Conductor th
 
 ### Format
 
-Every AskUserQuestion is a decision brief and must be sent as tool_use, not prose — unless the documented failure fallback above applies (interactive session + the call is unavailable/erroring), in which case the prose fallback is the correct output.
+Every AskUserQuestion decision has two parts: a markdown decision brief before the call, then a compact tool_use payload. Do not pack the full brief into the tool's `question` string. Every AskUserQuestion is a decision brief and must be sent as tool_use, not prose — unless the documented failure fallback above applies (interactive session + the call is unavailable/erroring), in which case the prose fallback is the correct output.
 
 ```
 D<N> — <one-line question title>
@@ -411,6 +411,12 @@ Effort both-scales: when an option involves effort, label both human-team and CC
 
 Net line closes the tradeoff. Per-skill instructions may add stricter rules.
 
+Tool payload rules:
+- `question` is only the decision prompt: one sentence, no newlines, <=80 chars.
+- Background, regrounding, ELI10, stakes, recommendation, pros/cons, and trade-off tables stay in the markdown brief before the tool call.
+- Ask one decision per tool call when possible; batch at most two related questions/tabs. Sequence independent decisions instead of sending 3+ tabs.
+- Do not duplicate the same trade-off text in both `question` and `options[].description`. Prefer putting option-specific trade-offs in `options[].description`.
+
 ### Handling 5+ options — split, never drop
 
 AskUserQuestion caps every call at **4 options**. With 5+ real options, NEVER
@@ -458,6 +464,10 @@ Before calling AskUserQuestion, verify:
 - [ ] (recommended) label on one option (even for neutral-posture)
 - [ ] Dual-scale effort labels on effort-bearing options (human / CC)
 - [ ] Net line closes the decision
+- [ ] `question` is one sentence, no newlines, <=80 chars
+- [ ] Tool call has no more than two related questions/tabs
+- [ ] No duplicated trade-off text between `question` and `options[].description`
+- [ ] You wrote the brief, then called the tool_use payload
 - [ ] You are calling the tool, not writing prose — unless `CONDUCTOR_SESSION: true` (then prose is the DEFAULT, not the tool) OR the documented failure fallback applies (then: prose with the mandatory triad — issue ELI10, per-choice Completeness, Recommendation + `(recommended)` — and a "reply with a letter" instruction, then STOP)
 - [ ] Non-ASCII characters (CJK / accents) written directly, NOT \u-escaped
 - [ ] If you had 5+ options, you split (or batched into ≤4-groups) — did NOT drop any

ios-clean/SKILL.md

@@ -341,7 +341,7 @@ Layout: a `D<N>` title + a one-line note to reply with a letter (in Conductor th
 
 ### Format
 
-Every AskUserQuestion is a decision brief and must be sent as tool_use, not prose — unless the documented failure fallback above applies (interactive session + the call is unavailable/erroring), in which case the prose fallback is the correct output.
+Every AskUserQuestion decision has two parts: a markdown decision brief before the call, then a compact tool_use payload. Do not pack the full brief into the tool's `question` string. Every AskUserQuestion is a decision brief and must be sent as tool_use, not prose — unless the documented failure fallback above applies (interactive session + the call is unavailable/erroring), in which case the prose fallback is the correct output.
 
 ```
 D<N> — <one-line question title>
@@ -374,6 +374,12 @@ Effort both-scales: when an option involves effort, label both human-team and CC
 
 Net line closes the tradeoff. Per-skill instructions may add stricter rules.
 
+Tool payload rules:
+- `question` is only the decision prompt: one sentence, no newlines, <=80 chars.
+- Background, regrounding, ELI10, stakes, recommendation, pros/cons, and trade-off tables stay in the markdown brief before the tool call.
+- Ask one decision per tool call when possible; batch at most two related questions/tabs. Sequence independent decisions instead of sending 3+ tabs.
+- Do not duplicate the same trade-off text in both `question` and `options[].description`. Prefer putting option-specific trade-offs in `options[].description`.
+
 ### Handling 5+ options — split, never drop
 
 AskUserQuestion caps every call at **4 options**. With 5+ real options, NEVER
@@ -421,6 +427,10 @@ Before calling AskUserQuestion, verify:
 - [ ] (recommended) label on one option (even for neutral-posture)
 - [ ] Dual-scale effort labels on effort-bearing options (human / CC)
 - [ ] Net line closes the decision
+- [ ] `question` is one sentence, no newlines, <=80 chars
+- [ ] Tool call has no more than two related questions/tabs
+- [ ] No duplicated trade-off text between `question` and `options[].description`
+- [ ] You wrote the brief, then called the tool_use payload
 - [ ] You are calling the tool, not writing prose — unless `CONDUCTOR_SESSION: true` (then prose is the DEFAULT, not the tool) OR the documented failure fallback applies (then: prose with the mandatory triad — issue ELI10, per-choice Completeness, Recommendation + `(recommended)` — and a "reply with a letter" instruction, then STOP)
 - [ ] Non-ASCII characters (CJK / accents) written directly, NOT \u-escaped
 - [ ] If you had 5+ options, you split (or batched into ≤4-groups) — did NOT drop any

ios-design-review/SKILL.md

@@ -343,7 +343,7 @@ Layout: a `D<N>` title + a one-line note to reply with a letter (in Conductor th
 
 ### Format
 
-Every AskUserQuestion is a decision brief and must be sent as tool_use, not prose — unless the documented failure fallback above applies (interactive session + the call is unavailable/erroring), in which case the prose fallback is the correct output.
+Every AskUserQuestion decision has two parts: a markdown decision brief before the call, then a compact tool_use payload. Do not pack the full brief into the tool's `question` string. Every AskUserQuestion is a decision brief and must be sent as tool_use, not prose — unless the documented failure fallback above applies (interactive session + the call is unavailable/erroring), in which case the prose fallback is the correct output.
 
 ```
 D<N> — <one-line question title>
@@ -376,6 +376,12 @@ Effort both-scales: when an option involves effort, label both human-team and CC
 
 Net line closes the tradeoff. Per-skill instructions may add stricter rules.
 
+Tool payload rules:
+- `question` is only the decision prompt: one sentence, no newlines, <=80 chars.
+- Background, regrounding, ELI10, stakes, recommendation, pros/cons, and trade-off tables stay in the markdown brief before the tool call.
+- Ask one decision per tool call when possible; batch at most two related questions/tabs. Sequence independent decisions instead of sending 3+ tabs.
+- Do not duplicate the same trade-off text in both `question` and `options[].description`. Prefer putting option-specific trade-offs in `options[].description`.
+
 ### Handling 5+ options — split, never drop
 
 AskUserQuestion caps every call at **4 options**. With 5+ real options, NEVER
@@ -423,6 +429,10 @@ Before calling AskUserQuestion, verify:
 - [ ] (recommended) label on one option (even for neutral-posture)
 - [ ] Dual-scale effort labels on effort-bearing options (human / CC)
 - [ ] Net line closes the decision
+- [ ] `question` is one sentence, no newlines, <=80 chars
+- [ ] Tool call has no more than two related questions/tabs
+- [ ] No duplicated trade-off text between `question` and `options[].description`
+- [ ] You wrote the brief, then called the tool_use payload
 - [ ] You are calling the tool, not writing prose — unless `CONDUCTOR_SESSION: true` (then prose is the DEFAULT, not the tool) OR the documented failure fallback applies (then: prose with the mandatory triad — issue ELI10, per-choice Completeness, Recommendation + `(recommended)` — and a "reply with a letter" instruction, then STOP)
 - [ ] Non-ASCII characters (CJK / accents) written directly, NOT \u-escaped
 - [ ] If you had 5+ options, you split (or batched into ≤4-groups) — did NOT drop any

ios-fix/SKILL.md

@@ -344,7 +344,7 @@ Layout: a `D<N>` title + a one-line note to reply with a letter (in Conductor th
 
 ### Format
 
-Every AskUserQuestion is a decision brief and must be sent as tool_use, not prose — unless the documented failure fallback above applies (interactive session + the call is unavailable/erroring), in which case the prose fallback is the correct output.
+Every AskUserQuestion decision has two parts: a markdown decision brief before the call, then a compact tool_use payload. Do not pack the full brief into the tool's `question` string. Every AskUserQuestion is a decision brief and must be sent as tool_use, not prose — unless the documented failure fallback above applies (interactive session + the call is unavailable/erroring), in which case the prose fallback is the correct output.
 
 ```
 D<N> — <one-line question title>
@@ -377,6 +377,12 @@ Effort both-scales: when an option involves effort, label both human-team and CC
 
 Net line closes the tradeoff. Per-skill instructions may add stricter rules.
 
+Tool payload rules:
+- `question` is only the decision prompt: one sentence, no newlines, <=80 chars.
+- Background, regrounding, ELI10, stakes, recommendation, pros/cons, and trade-off tables stay in the markdown brief before the tool call.
+- Ask one decision per tool call when possible; batch at most two related questions/tabs. Sequence independent decisions instead of sending 3+ tabs.
+- Do not duplicate the same trade-off text in both `question` and `options[].description`. Prefer putting option-specific trade-offs in `options[].description`.
+
 ### Handling 5+ options — split, never drop
 
 AskUserQuestion caps every call at **4 options**. With 5+ real options, NEVER
@@ -424,6 +430,10 @@ Before calling AskUserQuestion, verify:
 - [ ] (recommended) label on one option (even for neutral-posture)
 - [ ] Dual-scale effort labels on effort-bearing options (human / CC)
 - [ ] Net line closes the decision
+- [ ] `question` is one sentence, no newlines, <=80 chars
+- [ ] Tool call has no more than two related questions/tabs
+- [ ] No duplicated trade-off text between `question` and `options[].description`
+- [ ] You wrote the brief, then called the tool_use payload
 - [ ] You are calling the tool, not writing prose — unless `CONDUCTOR_SESSION: true` (then prose is the DEFAULT, not the tool) OR the documented failure fallback applies (then: prose with the mandatory triad — issue ELI10, per-choice Completeness, Recommendation + `(recommended)` — and a "reply with a letter" instruction, then STOP)
 - [ ] Non-ASCII characters (CJK / accents) written directly, NOT \u-escaped
 - [ ] If you had 5+ options, you split (or batched into ≤4-groups) — did NOT drop any

ios-qa/SKILL.md

@@ -347,7 +347,7 @@ Layout: a `D<N>` title + a one-line note to reply with a letter (in Conductor th
 
 ### Format
 
-Every AskUserQuestion is a decision brief and must be sent as tool_use, not prose — unless the documented failure fallback above applies (interactive session + the call is unavailable/erroring), in which case the prose fallback is the correct output.
+Every AskUserQuestion decision has two parts: a markdown decision brief before the call, then a compact tool_use payload. Do not pack the full brief into the tool's `question` string. Every AskUserQuestion is a decision brief and must be sent as tool_use, not prose — unless the documented failure fallback above applies (interactive session + the call is unavailable/erroring), in which case the prose fallback is the correct output.
 
 ```
 D<N> — <one-line question title>
@@ -380,6 +380,12 @@ Effort both-scales: when an option involves effort, label both human-team and CC
 
 Net line closes the tradeoff. Per-skill instructions may add stricter rules.
 
+Tool payload rules:
+- `question` is only the decision prompt: one sentence, no newlines, <=80 chars.
+- Background, regrounding, ELI10, stakes, recommendation, pros/cons, and trade-off tables stay in the markdown brief before the tool call.
+- Ask one decision per tool call when possible; batch at most two related questions/tabs. Sequence independent decisions instead of sending 3+ tabs.
+- Do not duplicate the same trade-off text in both `question` and `options[].description`. Prefer putting option-specific trade-offs in `options[].description`.
+
 ### Handling 5+ options — split, never drop
 
 AskUserQuestion caps every call at **4 options**. With 5+ real options, NEVER
@@ -427,6 +433,10 @@ Before calling AskUserQuestion, verify:
 - [ ] (recommended) label on one option (even for neutral-posture)
 - [ ] Dual-scale effort labels on effort-bearing options (human / CC)
 - [ ] Net line closes the decision
+- [ ] `question` is one sentence, no newlines, <=80 chars
+- [ ] Tool call has no more than two related questions/tabs
+- [ ] No duplicated trade-off text between `question` and `options[].description`
+- [ ] You wrote the brief, then called the tool_use payload
 - [ ] You are calling the tool, not writing prose — unless `CONDUCTOR_SESSION: true` (then prose is the DEFAULT, not the tool) OR the documented failure fallback applies (then: prose with the mandatory triad — issue ELI10, per-choice Completeness, Recommendation + `(recommended)` — and a "reply with a letter" instruction, then STOP)
 - [ ] Non-ASCII characters (CJK / accents) written directly, NOT \u-escaped
 - [ ] If you had 5+ options, you split (or batched into ≤4-groups) — did NOT drop any

ios-sync/SKILL.md

@@ -341,7 +341,7 @@ Layout: a `D<N>` title + a one-line note to reply with a letter (in Conductor th
 
 ### Format
 
-Every AskUserQuestion is a decision brief and must be sent as tool_use, not prose — unless the documented failure fallback above applies (interactive session + the call is unavailable/erroring), in which case the prose fallback is the correct output.
+Every AskUserQuestion decision has two parts: a markdown decision brief before the call, then a compact tool_use payload. Do not pack the full brief into the tool's `question` string. Every AskUserQuestion is a decision brief and must be sent as tool_use, not prose — unless the documented failure fallback above applies (interactive session + the call is unavailable/erroring), in which case the prose fallback is the correct output.
 
 ```
 D<N> — <one-line question title>
@@ -374,6 +374,12 @@ Effort both-scales: when an option involves effort, label both human-team and CC
 
 Net line closes the tradeoff. Per-skill instructions may add stricter rules.
 
+Tool payload rules:
+- `question` is only the decision prompt: one sentence, no newlines, <=80 chars.
+- Background, regrounding, ELI10, stakes, recommendation, pros/cons, and trade-off tables stay in the markdown brief before the tool call.
+- Ask one decision per tool call when possible; batch at most two related questions/tabs. Sequence independent decisions instead of sending 3+ tabs.
+- Do not duplicate the same trade-off text in both `question` and `options[].description`. Prefer putting option-specific trade-offs in `options[].description`.
+
 ### Handling 5+ options — split, never drop
 
 AskUserQuestion caps every call at **4 options**. With 5+ real options, NEVER
@@ -421,6 +427,10 @@ Before calling AskUserQuestion, verify:
 - [ ] (recommended) label on one option (even for neutral-posture)
 - [ ] Dual-scale effort labels on effort-bearing options (human / CC)
 - [ ] Net line closes the decision
+- [ ] `question` is one sentence, no newlines, <=80 chars
+- [ ] Tool call has no more than two related questions/tabs
+- [ ] No duplicated trade-off text between `question` and `options[].description`
+- [ ] You wrote the brief, then called the tool_use payload
 - [ ] You are calling the tool, not writing prose — unless `CONDUCTOR_SESSION: true` (then prose is the DEFAULT, not the tool) OR the documented failure fallback applies (then: prose with the mandatory triad — issue ELI10, per-choice Completeness, Recommendation + `(recommended)` — and a "reply with a letter" instruction, then STOP)
 - [ ] Non-ASCII characters (CJK / accents) written directly, NOT \u-escaped
 - [ ] If you had 5+ options, you split (or batched into ≤4-groups) — did NOT drop any

land-and-deploy/SKILL.md

@@ -336,7 +336,7 @@ Layout: a `D<N>` title + a one-line note to reply with a letter (in Conductor th
 
 ### Format
 
-Every AskUserQuestion is a decision brief and must be sent as tool_use, not prose — unless the documented failure fallback above applies (interactive session + the call is unavailable/erroring), in which case the prose fallback is the correct output.
+Every AskUserQuestion decision has two parts: a markdown decision brief before the call, then a compact tool_use payload. Do not pack the full brief into the tool's `question` string. Every AskUserQuestion is a decision brief and must be sent as tool_use, not prose — unless the documented failure fallback above applies (interactive session + the call is unavailable/erroring), in which case the prose fallback is the correct output.
 
 ```
 D<N> — <one-line question title>
@@ -369,6 +369,12 @@ Effort both-scales: when an option involves effort, label both human-team and CC
 
 Net line closes the tradeoff. Per-skill instructions may add stricter rules.
 
+Tool payload rules:
+- `question` is only the decision prompt: one sentence, no newlines, <=80 chars.
+- Background, regrounding, ELI10, stakes, recommendation, pros/cons, and trade-off tables stay in the markdown brief before the tool call.
+- Ask one decision per tool call when possible; batch at most two related questions/tabs. Sequence independent decisions instead of sending 3+ tabs.
+- Do not duplicate the same trade-off text in both `question` and `options[].description`. Prefer putting option-specific trade-offs in `options[].description`.
+
 ### Handling 5+ options — split, never drop
 
 AskUserQuestion caps every call at **4 options**. With 5+ real options, NEVER
@@ -416,6 +422,10 @@ Before calling AskUserQuestion, verify:
 - [ ] (recommended) label on one option (even for neutral-posture)
 - [ ] Dual-scale effort labels on effort-bearing options (human / CC)
 - [ ] Net line closes the decision
+- [ ] `question` is one sentence, no newlines, <=80 chars
+- [ ] Tool call has no more than two related questions/tabs
+- [ ] No duplicated trade-off text between `question` and `options[].description`
+- [ ] You wrote the brief, then called the tool_use payload
 - [ ] You are calling the tool, not writing prose — unless `CONDUCTOR_SESSION: true` (then prose is the DEFAULT, not the tool) OR the documented failure fallback applies (then: prose with the mandatory triad — issue ELI10, per-choice Completeness, Recommendation + `(recommended)` — and a "reply with a letter" instruction, then STOP)
 - [ ] Non-ASCII characters (CJK / accents) written directly, NOT \u-escaped
 - [ ] If you had 5+ options, you split (or batched into ≤4-groups) — did NOT drop any

landing-report/SKILL.md

@@ -337,7 +337,7 @@ Layout: a `D<N>` title + a one-line note to reply with a letter (in Conductor th
 
 ### Format
 
-Every AskUserQuestion is a decision brief and must be sent as tool_use, not prose — unless the documented failure fallback above applies (interactive session + the call is unavailable/erroring), in which case the prose fallback is the correct output.
+Every AskUserQuestion decision has two parts: a markdown decision brief before the call, then a compact tool_use payload. Do not pack the full brief into the tool's `question` string. Every AskUserQuestion is a decision brief and must be sent as tool_use, not prose — unless the documented failure fallback above applies (interactive session + the call is unavailable/erroring), in which case the prose fallback is the correct output.
 
 ```
 D<N> — <one-line question title>
@@ -370,6 +370,12 @@ Effort both-scales: when an option involves effort, label both human-team and CC
 
 Net line closes the tradeoff. Per-skill instructions may add stricter rules.
 
+Tool payload rules:
+- `question` is only the decision prompt: one sentence, no newlines, <=80 chars.
+- Background, regrounding, ELI10, stakes, recommendation, pros/cons, and trade-off tables stay in the markdown brief before the tool call.
+- Ask one decision per tool call when possible; batch at most two related questions/tabs. Sequence independent decisions instead of sending 3+ tabs.
+- Do not duplicate the same trade-off text in both `question` and `options[].description`. Prefer putting option-specific trade-offs in `options[].description`.
+
 ### Handling 5+ options — split, never drop
 
 AskUserQuestion caps every call at **4 options**. With 5+ real options, NEVER
@@ -417,6 +423,10 @@ Before calling AskUserQuestion, verify:
 - [ ] (recommended) label on one option (even for neutral-posture)
 - [ ] Dual-scale effort labels on effort-bearing options (human / CC)
 - [ ] Net line closes the decision
+- [ ] `question` is one sentence, no newlines, <=80 chars
+- [ ] Tool call has no more than two related questions/tabs
+- [ ] No duplicated trade-off text between `question` and `options[].description`
+- [ ] You wrote the brief, then called the tool_use payload
 - [ ] You are calling the tool, not writing prose — unless `CONDUCTOR_SESSION: true` (then prose is the DEFAULT, not the tool) OR the documented failure fallback applies (then: prose with the mandatory triad — issue ELI10, per-choice Completeness, Recommendation + `(recommended)` — and a "reply with a letter" instruction, then STOP)
 - [ ] Non-ASCII characters (CJK / accents) written directly, NOT \u-escaped
 - [ ] If you had 5+ options, you split (or batched into ≤4-groups) — did NOT drop any

learn/SKILL.md

@@ -339,7 +339,7 @@ Layout: a `D<N>` title + a one-line note to reply with a letter (in Conductor th
 
 ### Format
 
-Every AskUserQuestion is a decision brief and must be sent as tool_use, not prose — unless the documented failure fallback above applies (interactive session + the call is unavailable/erroring), in which case the prose fallback is the correct output.
+Every AskUserQuestion decision has two parts: a markdown decision brief before the call, then a compact tool_use payload. Do not pack the full brief into the tool's `question` string. Every AskUserQuestion is a decision brief and must be sent as tool_use, not prose — unless the documented failure fallback above applies (interactive session + the call is unavailable/erroring), in which case the prose fallback is the correct output.
 
 ```
 D<N> — <one-line question title>
@@ -372,6 +372,12 @@ Effort both-scales: when an option involves effort, label both human-team and CC
 
 Net line closes the tradeoff. Per-skill instructions may add stricter rules.
 
+Tool payload rules:
+- `question` is only the decision prompt: one sentence, no newlines, <=80 chars.
+- Background, regrounding, ELI10, stakes, recommendation, pros/cons, and trade-off tables stay in the markdown brief before the tool call.
+- Ask one decision per tool call when possible; batch at most two related questions/tabs. Sequence independent decisions instead of sending 3+ tabs.
+- Do not duplicate the same trade-off text in both `question` and `options[].description`. Prefer putting option-specific trade-offs in `options[].description`.
+
 ### Handling 5+ options — split, never drop
 
 AskUserQuestion caps every call at **4 options**. With 5+ real options, NEVER
@@ -419,6 +425,10 @@ Before calling AskUserQuestion, verify:
 - [ ] (recommended) label on one option (even for neutral-posture)
 - [ ] Dual-scale effort labels on effort-bearing options (human / CC)
 - [ ] Net line closes the decision
+- [ ] `question` is one sentence, no newlines, <=80 chars
+- [ ] Tool call has no more than two related questions/tabs
+- [ ] No duplicated trade-off text between `question` and `options[].description`
+- [ ] You wrote the brief, then called the tool_use payload
 - [ ] You are calling the tool, not writing prose — unless `CONDUCTOR_SESSION: true` (then prose is the DEFAULT, not the tool) OR the documented failure fallback applies (then: prose with the mandatory triad — issue ELI10, per-choice Completeness, Recommendation + `(recommended)` — and a "reply with a letter" instruction, then STOP)
 - [ ] Non-ASCII characters (CJK / accents) written directly, NOT \u-escaped
 - [ ] If you had 5+ options, you split (or batched into ≤4-groups) — did NOT drop any

office-hours/SKILL.md

@@ -374,7 +374,7 @@ Layout: a `D<N>` title + a one-line note to reply with a letter (in Conductor th
 
 ### Format
 
-Every AskUserQuestion is a decision brief and must be sent as tool_use, not prose — unless the documented failure fallback above applies (interactive session + the call is unavailable/erroring), in which case the prose fallback is the correct output.
+Every AskUserQuestion decision has two parts: a markdown decision brief before the call, then a compact tool_use payload. Do not pack the full brief into the tool's `question` string. Every AskUserQuestion is a decision brief and must be sent as tool_use, not prose — unless the documented failure fallback above applies (interactive session + the call is unavailable/erroring), in which case the prose fallback is the correct output.
 
 ```
 D<N> — <one-line question title>
@@ -407,6 +407,12 @@ Effort both-scales: when an option involves effort, label both human-team and CC
 
 Net line closes the tradeoff. Per-skill instructions may add stricter rules.
 
+Tool payload rules:
+- `question` is only the decision prompt: one sentence, no newlines, <=80 chars.
+- Background, regrounding, ELI10, stakes, recommendation, pros/cons, and trade-off tables stay in the markdown brief before the tool call.
+- Ask one decision per tool call when possible; batch at most two related questions/tabs. Sequence independent decisions instead of sending 3+ tabs.
+- Do not duplicate the same trade-off text in both `question` and `options[].description`. Prefer putting option-specific trade-offs in `options[].description`.
+
 ### Handling 5+ options — split, never drop
 
 AskUserQuestion caps every call at **4 options**. With 5+ real options, NEVER
@@ -454,6 +460,10 @@ Before calling AskUserQuestion, verify:
 - [ ] (recommended) label on one option (even for neutral-posture)
 - [ ] Dual-scale effort labels on effort-bearing options (human / CC)
 - [ ] Net line closes the decision
+- [ ] `question` is one sentence, no newlines, <=80 chars
+- [ ] Tool call has no more than two related questions/tabs
+- [ ] No duplicated trade-off text between `question` and `options[].description`
+- [ ] You wrote the brief, then called the tool_use payload
 - [ ] You are calling the tool, not writing prose — unless `CONDUCTOR_SESSION: true` (then prose is the DEFAULT, not the tool) OR the documented failure fallback applies (then: prose with the mandatory triad — issue ELI10, per-choice Completeness, Recommendation + `(recommended)` — and a "reply with a letter" instruction, then STOP)
 - [ ] Non-ASCII characters (CJK / accents) written directly, NOT \u-escaped
 - [ ] If you had 5+ options, you split (or batched into ≤4-groups) — did NOT drop any

open-gstack-browser/SKILL.md

@@ -336,7 +336,7 @@ Layout: a `D<N>` title + a one-line note to reply with a letter (in Conductor th
 
 ### Format
 
-Every AskUserQuestion is a decision brief and must be sent as tool_use, not prose — unless the documented failure fallback above applies (interactive session + the call is unavailable/erroring), in which case the prose fallback is the correct output.
+Every AskUserQuestion decision has two parts: a markdown decision brief before the call, then a compact tool_use payload. Do not pack the full brief into the tool's `question` string. Every AskUserQuestion is a decision brief and must be sent as tool_use, not prose — unless the documented failure fallback above applies (interactive session + the call is unavailable/erroring), in which case the prose fallback is the correct output.
 
 ```
 D<N> — <one-line question title>
@@ -369,6 +369,12 @@ Effort both-scales: when an option involves effort, label both human-team and CC
 
 Net line closes the tradeoff. Per-skill instructions may add stricter rules.
 
+Tool payload rules:
+- `question` is only the decision prompt: one sentence, no newlines, <=80 chars.
+- Background, regrounding, ELI10, stakes, recommendation, pros/cons, and trade-off tables stay in the markdown brief before the tool call.
+- Ask one decision per tool call when possible; batch at most two related questions/tabs. Sequence independent decisions instead of sending 3+ tabs.
+- Do not duplicate the same trade-off text in both `question` and `options[].description`. Prefer putting option-specific trade-offs in `options[].description`.
+
 ### Handling 5+ options — split, never drop
 
 AskUserQuestion caps every call at **4 options**. With 5+ real options, NEVER
@@ -416,6 +422,10 @@ Before calling AskUserQuestion, verify:
 - [ ] (recommended) label on one option (even for neutral-posture)
 - [ ] Dual-scale effort labels on effort-bearing options (human / CC)
 - [ ] Net line closes the decision
+- [ ] `question` is one sentence, no newlines, <=80 chars
+- [ ] Tool call has no more than two related questions/tabs
+- [ ] No duplicated trade-off text between `question` and `options[].description`
+- [ ] You wrote the brief, then called the tool_use payload
 - [ ] You are calling the tool, not writing prose — unless `CONDUCTOR_SESSION: true` (then prose is the DEFAULT, not the tool) OR the documented failure fallback applies (then: prose with the mandatory triad — issue ELI10, per-choice Completeness, Recommendation + `(recommended)` — and a "reply with a letter" instruction, then STOP)
 - [ ] Non-ASCII characters (CJK / accents) written directly, NOT \u-escaped
 - [ ] If you had 5+ options, you split (or batched into ≤4-groups) — did NOT drop any

pair-agent/SKILL.md

@@ -338,7 +338,7 @@ Layout: a `D<N>` title + a one-line note to reply with a letter (in Conductor th
 
 ### Format
 
-Every AskUserQuestion is a decision brief and must be sent as tool_use, not prose — unless the documented failure fallback above applies (interactive session + the call is unavailable/erroring), in which case the prose fallback is the correct output.
+Every AskUserQuestion decision has two parts: a markdown decision brief before the call, then a compact tool_use payload. Do not pack the full brief into the tool's `question` string. Every AskUserQuestion is a decision brief and must be sent as tool_use, not prose — unless the documented failure fallback above applies (interactive session + the call is unavailable/erroring), in which case the prose fallback is the correct output.
 
 ```
 D<N> — <one-line question title>
@@ -371,6 +371,12 @@ Effort both-scales: when an option involves effort, label both human-team and CC
 
 Net line closes the tradeoff. Per-skill instructions may add stricter rules.
 
+Tool payload rules:
+- `question` is only the decision prompt: one sentence, no newlines, <=80 chars.
+- Background, regrounding, ELI10, stakes, recommendation, pros/cons, and trade-off tables stay in the markdown brief before the tool call.
+- Ask one decision per tool call when possible; batch at most two related questions/tabs. Sequence independent decisions instead of sending 3+ tabs.
+- Do not duplicate the same trade-off text in both `question` and `options[].description`. Prefer putting option-specific trade-offs in `options[].description`.
+
 ### Handling 5+ options — split, never drop
 
 AskUserQuestion caps every call at **4 options**. With 5+ real options, NEVER
@@ -418,6 +424,10 @@ Before calling AskUserQuestion, verify:
 - [ ] (recommended) label on one option (even for neutral-posture)
 - [ ] Dual-scale effort labels on effort-bearing options (human / CC)
 - [ ] Net line closes the decision
+- [ ] `question` is one sentence, no newlines, <=80 chars
+- [ ] Tool call has no more than two related questions/tabs
+- [ ] No duplicated trade-off text between `question` and `options[].description`
+- [ ] You wrote the brief, then called the tool_use payload
 - [ ] You are calling the tool, not writing prose — unless `CONDUCTOR_SESSION: true` (then prose is the DEFAULT, not the tool) OR the documented failure fallback applies (then: prose with the mandatory triad — issue ELI10, per-choice Completeness, Recommendation + `(recommended)` — and a "reply with a letter" instruction, then STOP)
 - [ ] Non-ASCII characters (CJK / accents) written directly, NOT \u-escaped
 - [ ] If you had 5+ options, you split (or batched into ≤4-groups) — did NOT drop any

plan-ceo-review/SKILL.md

@@ -368,7 +368,7 @@ Layout: a `D<N>` title + a one-line note to reply with a letter (in Conductor th
 
 ### Format
 
-Every AskUserQuestion is a decision brief and must be sent as tool_use, not prose — unless the documented failure fallback above applies (interactive session + the call is unavailable/erroring), in which case the prose fallback is the correct output.
+Every AskUserQuestion decision has two parts: a markdown decision brief before the call, then a compact tool_use payload. Do not pack the full brief into the tool's `question` string. Every AskUserQuestion is a decision brief and must be sent as tool_use, not prose — unless the documented failure fallback above applies (interactive session + the call is unavailable/erroring), in which case the prose fallback is the correct output.
 
 ```
 D<N> — <one-line question title>
@@ -401,6 +401,12 @@ Effort both-scales: when an option involves effort, label both human-team and CC
 
 Net line closes the tradeoff. Per-skill instructions may add stricter rules.
 
+Tool payload rules:
+- `question` is only the decision prompt: one sentence, no newlines, <=80 chars.
+- Background, regrounding, ELI10, stakes, recommendation, pros/cons, and trade-off tables stay in the markdown brief before the tool call.
+- Ask one decision per tool call when possible; batch at most two related questions/tabs. Sequence independent decisions instead of sending 3+ tabs.
+- Do not duplicate the same trade-off text in both `question` and `options[].description`. Prefer putting option-specific trade-offs in `options[].description`.
+
 ### Handling 5+ options — split, never drop
 
 AskUserQuestion caps every call at **4 options**. With 5+ real options, NEVER
@@ -448,6 +454,10 @@ Before calling AskUserQuestion, verify:
 - [ ] (recommended) label on one option (even for neutral-posture)
 - [ ] Dual-scale effort labels on effort-bearing options (human / CC)
 - [ ] Net line closes the decision
+- [ ] `question` is one sentence, no newlines, <=80 chars
+- [ ] Tool call has no more than two related questions/tabs
+- [ ] No duplicated trade-off text between `question` and `options[].description`
+- [ ] You wrote the brief, then called the tool_use payload
 - [ ] You are calling the tool, not writing prose — unless `CONDUCTOR_SESSION: true` (then prose is the DEFAULT, not the tool) OR the documented failure fallback applies (then: prose with the mandatory triad — issue ELI10, per-choice Completeness, Recommendation + `(recommended)` — and a "reply with a letter" instruction, then STOP)
 - [ ] Non-ASCII characters (CJK / accents) written directly, NOT \u-escaped
 - [ ] If you had 5+ options, you split (or batched into ≤4-groups) — did NOT drop any

plan-design-review/SKILL.md

@@ -340,7 +340,7 @@ Layout: a `D<N>` title + a one-line note to reply with a letter (in Conductor th
 
 ### Format
 
-Every AskUserQuestion is a decision brief and must be sent as tool_use, not prose — unless the documented failure fallback above applies (interactive session + the call is unavailable/erroring), in which case the prose fallback is the correct output.
+Every AskUserQuestion decision has two parts: a markdown decision brief before the call, then a compact tool_use payload. Do not pack the full brief into the tool's `question` string. Every AskUserQuestion is a decision brief and must be sent as tool_use, not prose — unless the documented failure fallback above applies (interactive session + the call is unavailable/erroring), in which case the prose fallback is the correct output.
 
 ```
 D<N> — <one-line question title>
@@ -373,6 +373,12 @@ Effort both-scales: when an option involves effort, label both human-team and CC
 
 Net line closes the tradeoff. Per-skill instructions may add stricter rules.
 
+Tool payload rules:
+- `question` is only the decision prompt: one sentence, no newlines, <=80 chars.
+- Background, regrounding, ELI10, stakes, recommendation, pros/cons, and trade-off tables stay in the markdown brief before the tool call.
+- Ask one decision per tool call when possible; batch at most two related questions/tabs. Sequence independent decisions instead of sending 3+ tabs.
+- Do not duplicate the same trade-off text in both `question` and `options[].description`. Prefer putting option-specific trade-offs in `options[].description`.
+
 ### Handling 5+ options — split, never drop
 
 AskUserQuestion caps every call at **4 options**. With 5+ real options, NEVER
@@ -420,6 +426,10 @@ Before calling AskUserQuestion, verify:
 - [ ] (recommended) label on one option (even for neutral-posture)
 - [ ] Dual-scale effort labels on effort-bearing options (human / CC)
 - [ ] Net line closes the decision
+- [ ] `question` is one sentence, no newlines, <=80 chars
+- [ ] Tool call has no more than two related questions/tabs
+- [ ] No duplicated trade-off text between `question` and `options[].description`
+- [ ] You wrote the brief, then called the tool_use payload
 - [ ] You are calling the tool, not writing prose — unless `CONDUCTOR_SESSION: true` (then prose is the DEFAULT, not the tool) OR the documented failure fallback applies (then: prose with the mandatory triad — issue ELI10, per-choice Completeness, Recommendation + `(recommended)` — and a "reply with a letter" instruction, then STOP)
 - [ ] Non-ASCII characters (CJK / accents) written directly, NOT \u-escaped
 - [ ] If you had 5+ options, you split (or batched into ≤4-groups) — did NOT drop any

plan-devex-review/SKILL.md

@@ -346,7 +346,7 @@ Layout: a `D<N>` title + a one-line note to reply with a letter (in Conductor th
 
 ### Format
 
-Every AskUserQuestion is a decision brief and must be sent as tool_use, not prose — unless the documented failure fallback above applies (interactive session + the call is unavailable/erroring), in which case the prose fallback is the correct output.
+Every AskUserQuestion decision has two parts: a markdown decision brief before the call, then a compact tool_use payload. Do not pack the full brief into the tool's `question` string. Every AskUserQuestion is a decision brief and must be sent as tool_use, not prose — unless the documented failure fallback above applies (interactive session + the call is unavailable/erroring), in which case the prose fallback is the correct output.
 
 ```
 D<N> — <one-line question title>
@@ -379,6 +379,12 @@ Effort both-scales: when an option involves effort, label both human-team and CC
 
 Net line closes the tradeoff. Per-skill instructions may add stricter rules.
 
+Tool payload rules:
+- `question` is only the decision prompt: one sentence, no newlines, <=80 chars.
+- Background, regrounding, ELI10, stakes, recommendation, pros/cons, and trade-off tables stay in the markdown brief before the tool call.
+- Ask one decision per tool call when possible; batch at most two related questions/tabs. Sequence independent decisions instead of sending 3+ tabs.
+- Do not duplicate the same trade-off text in both `question` and `options[].description`. Prefer putting option-specific trade-offs in `options[].description`.
+
 ### Handling 5+ options — split, never drop
 
 AskUserQuestion caps every call at **4 options**. With 5+ real options, NEVER
@@ -426,6 +432,10 @@ Before calling AskUserQuestion, verify:
 - [ ] (recommended) label on one option (even for neutral-posture)
 - [ ] Dual-scale effort labels on effort-bearing options (human / CC)
 - [ ] Net line closes the decision
+- [ ] `question` is one sentence, no newlines, <=80 chars
+- [ ] Tool call has no more than two related questions/tabs
+- [ ] No duplicated trade-off text between `question` and `options[].description`
+- [ ] You wrote the brief, then called the tool_use payload
 - [ ] You are calling the tool, not writing prose — unless `CONDUCTOR_SESSION: true` (then prose is the DEFAULT, not the tool) OR the documented failure fallback applies (then: prose with the mandatory triad — issue ELI10, per-choice Completeness, Recommendation + `(recommended)` — and a "reply with a letter" instruction, then STOP)
 - [ ] Non-ASCII characters (CJK / accents) written directly, NOT \u-escaped
 - [ ] If you had 5+ options, you split (or batched into ≤4-groups) — did NOT drop any

plan-eng-review/SKILL.md

@@ -344,7 +344,7 @@ Layout: a `D<N>` title + a one-line note to reply with a letter (in Conductor th
 
 ### Format
 
-Every AskUserQuestion is a decision brief and must be sent as tool_use, not prose — unless the documented failure fallback above applies (interactive session + the call is unavailable/erroring), in which case the prose fallback is the correct output.
+Every AskUserQuestion decision has two parts: a markdown decision brief before the call, then a compact tool_use payload. Do not pack the full brief into the tool's `question` string. Every AskUserQuestion is a decision brief and must be sent as tool_use, not prose — unless the documented failure fallback above applies (interactive session + the call is unavailable/erroring), in which case the prose fallback is the correct output.
 
 ```
 D<N> — <one-line question title>
@@ -377,6 +377,12 @@ Effort both-scales: when an option involves effort, label both human-team and CC
 
 Net line closes the tradeoff. Per-skill instructions may add stricter rules.
 
+Tool payload rules:
+- `question` is only the decision prompt: one sentence, no newlines, <=80 chars.
+- Background, regrounding, ELI10, stakes, recommendation, pros/cons, and trade-off tables stay in the markdown brief before the tool call.
+- Ask one decision per tool call when possible; batch at most two related questions/tabs. Sequence independent decisions instead of sending 3+ tabs.
+- Do not duplicate the same trade-off text in both `question` and `options[].description`. Prefer putting option-specific trade-offs in `options[].description`.
+
 ### Handling 5+ options — split, never drop
 
 AskUserQuestion caps every call at **4 options**. With 5+ real options, NEVER
@@ -424,6 +430,10 @@ Before calling AskUserQuestion, verify:
 - [ ] (recommended) label on one option (even for neutral-posture)
 - [ ] Dual-scale effort labels on effort-bearing options (human / CC)
 - [ ] Net line closes the decision
+- [ ] `question` is one sentence, no newlines, <=80 chars
+- [ ] Tool call has no more than two related questions/tabs
+- [ ] No duplicated trade-off text between `question` and `options[].description`
+- [ ] You wrote the brief, then called the tool_use payload
 - [ ] You are calling the tool, not writing prose — unless `CONDUCTOR_SESSION: true` (then prose is the DEFAULT, not the tool) OR the documented failure fallback applies (then: prose with the mandatory triad — issue ELI10, per-choice Completeness, Recommendation + `(recommended)` — and a "reply with a letter" instruction, then STOP)
 - [ ] Non-ASCII characters (CJK / accents) written directly, NOT \u-escaped
 - [ ] If you had 5+ options, you split (or batched into ≤4-groups) — did NOT drop any

plan-tune/SKILL.md

@@ -349,7 +349,7 @@ Layout: a `D<N>` title + a one-line note to reply with a letter (in Conductor th
 
 ### Format
 
-Every AskUserQuestion is a decision brief and must be sent as tool_use, not prose — unless the documented failure fallback above applies (interactive session + the call is unavailable/erroring), in which case the prose fallback is the correct output.
+Every AskUserQuestion decision has two parts: a markdown decision brief before the call, then a compact tool_use payload. Do not pack the full brief into the tool's `question` string. Every AskUserQuestion is a decision brief and must be sent as tool_use, not prose — unless the documented failure fallback above applies (interactive session + the call is unavailable/erroring), in which case the prose fallback is the correct output.
 
 ```
 D<N> — <one-line question title>
@@ -382,6 +382,12 @@ Effort both-scales: when an option involves effort, label both human-team and CC
 
 Net line closes the tradeoff. Per-skill instructions may add stricter rules.
 
+Tool payload rules:
+- `question` is only the decision prompt: one sentence, no newlines, <=80 chars.
+- Background, regrounding, ELI10, stakes, recommendation, pros/cons, and trade-off tables stay in the markdown brief before the tool call.
+- Ask one decision per tool call when possible; batch at most two related questions/tabs. Sequence independent decisions instead of sending 3+ tabs.
+- Do not duplicate the same trade-off text in both `question` and `options[].description`. Prefer putting option-specific trade-offs in `options[].description`.
+
 ### Handling 5+ options — split, never drop
 
 AskUserQuestion caps every call at **4 options**. With 5+ real options, NEVER
@@ -429,6 +435,10 @@ Before calling AskUserQuestion, verify:
 - [ ] (recommended) label on one option (even for neutral-posture)
 - [ ] Dual-scale effort labels on effort-bearing options (human / CC)
 - [ ] Net line closes the decision
+- [ ] `question` is one sentence, no newlines, <=80 chars
+- [ ] Tool call has no more than two related questions/tabs
+- [ ] No duplicated trade-off text between `question` and `options[].description`
+- [ ] You wrote the brief, then called the tool_use payload
 - [ ] You are calling the tool, not writing prose — unless `CONDUCTOR_SESSION: true` (then prose is the DEFAULT, not the tool) OR the documented failure fallback applies (then: prose with the mandatory triad — issue ELI10, per-choice Completeness, Recommendation + `(recommended)` — and a "reply with a letter" instruction, then STOP)
 - [ ] Non-ASCII characters (CJK / accents) written directly, NOT \u-escaped
 - [ ] If you had 5+ options, you split (or batched into ≤4-groups) — did NOT drop any

qa-only/SKILL.md

@@ -339,7 +339,7 @@ Layout: a `D<N>` title + a one-line note to reply with a letter (in Conductor th
 
 ### Format
 
-Every AskUserQuestion is a decision brief and must be sent as tool_use, not prose — unless the documented failure fallback above applies (interactive session + the call is unavailable/erroring), in which case the prose fallback is the correct output.
+Every AskUserQuestion decision has two parts: a markdown decision brief before the call, then a compact tool_use payload. Do not pack the full brief into the tool's `question` string. Every AskUserQuestion is a decision brief and must be sent as tool_use, not prose — unless the documented failure fallback above applies (interactive session + the call is unavailable/erroring), in which case the prose fallback is the correct output.
 
 ```
 D<N> — <one-line question title>
@@ -372,6 +372,12 @@ Effort both-scales: when an option involves effort, label both human-team and CC
 
 Net line closes the tradeoff. Per-skill instructions may add stricter rules.
 
+Tool payload rules:
+- `question` is only the decision prompt: one sentence, no newlines, <=80 chars.
+- Background, regrounding, ELI10, stakes, recommendation, pros/cons, and trade-off tables stay in the markdown brief before the tool call.
+- Ask one decision per tool call when possible; batch at most two related questions/tabs. Sequence independent decisions instead of sending 3+ tabs.
+- Do not duplicate the same trade-off text in both `question` and `options[].description`. Prefer putting option-specific trade-offs in `options[].description`.
+
 ### Handling 5+ options — split, never drop
 
 AskUserQuestion caps every call at **4 options**. With 5+ real options, NEVER
@@ -419,6 +425,10 @@ Before calling AskUserQuestion, verify:
 - [ ] (recommended) label on one option (even for neutral-posture)
 - [ ] Dual-scale effort labels on effort-bearing options (human / CC)
 - [ ] Net line closes the decision
+- [ ] `question` is one sentence, no newlines, <=80 chars
+- [ ] Tool call has no more than two related questions/tabs
+- [ ] No duplicated trade-off text between `question` and `options[].description`
+- [ ] You wrote the brief, then called the tool_use payload
 - [ ] You are calling the tool, not writing prose — unless `CONDUCTOR_SESSION: true` (then prose is the DEFAULT, not the tool) OR the documented failure fallback applies (then: prose with the mandatory triad — issue ELI10, per-choice Completeness, Recommendation + `(recommended)` — and a "reply with a letter" instruction, then STOP)
 - [ ] Non-ASCII characters (CJK / accents) written directly, NOT \u-escaped
 - [ ] If you had 5+ options, you split (or batched into ≤4-groups) — did NOT drop any

qa/SKILL.md

@@ -345,7 +345,7 @@ Layout: a `D<N>` title + a one-line note to reply with a letter (in Conductor th
 
 ### Format
 
-Every AskUserQuestion is a decision brief and must be sent as tool_use, not prose — unless the documented failure fallback above applies (interactive session + the call is unavailable/erroring), in which case the prose fallback is the correct output.
+Every AskUserQuestion decision has two parts: a markdown decision brief before the call, then a compact tool_use payload. Do not pack the full brief into the tool's `question` string. Every AskUserQuestion is a decision brief and must be sent as tool_use, not prose — unless the documented failure fallback above applies (interactive session + the call is unavailable/erroring), in which case the prose fallback is the correct output.
 
 ```
 D<N> — <one-line question title>
@@ -378,6 +378,12 @@ Effort both-scales: when an option involves effort, label both human-team and CC
 
 Net line closes the tradeoff. Per-skill instructions may add stricter rules.
 
+Tool payload rules:
+- `question` is only the decision prompt: one sentence, no newlines, <=80 chars.
+- Background, regrounding, ELI10, stakes, recommendation, pros/cons, and trade-off tables stay in the markdown brief before the tool call.
+- Ask one decision per tool call when possible; batch at most two related questions/tabs. Sequence independent decisions instead of sending 3+ tabs.
+- Do not duplicate the same trade-off text in both `question` and `options[].description`. Prefer putting option-specific trade-offs in `options[].description`.
+
 ### Handling 5+ options — split, never drop
 
 AskUserQuestion caps every call at **4 options**. With 5+ real options, NEVER
@@ -425,6 +431,10 @@ Before calling AskUserQuestion, verify:
 - [ ] (recommended) label on one option (even for neutral-posture)
 - [ ] Dual-scale effort labels on effort-bearing options (human / CC)
 - [ ] Net line closes the decision
+- [ ] `question` is one sentence, no newlines, <=80 chars
+- [ ] Tool call has no more than two related questions/tabs
+- [ ] No duplicated trade-off text between `question` and `options[].description`
+- [ ] You wrote the brief, then called the tool_use payload
 - [ ] You are calling the tool, not writing prose — unless `CONDUCTOR_SESSION: true` (then prose is the DEFAULT, not the tool) OR the documented failure fallback applies (then: prose with the mandatory triad — issue ELI10, per-choice Completeness, Recommendation + `(recommended)` — and a "reply with a letter" instruction, then STOP)
 - [ ] Non-ASCII characters (CJK / accents) written directly, NOT \u-escaped
 - [ ] If you had 5+ options, you split (or batched into ≤4-groups) — did NOT drop any

retro/SKILL.md

@@ -356,7 +356,7 @@ Layout: a `D<N>` title + a one-line note to reply with a letter (in Conductor th
 
 ### Format
 
-Every AskUserQuestion is a decision brief and must be sent as tool_use, not prose — unless the documented failure fallback above applies (interactive session + the call is unavailable/erroring), in which case the prose fallback is the correct output.
+Every AskUserQuestion decision has two parts: a markdown decision brief before the call, then a compact tool_use payload. Do not pack the full brief into the tool's `question` string. Every AskUserQuestion is a decision brief and must be sent as tool_use, not prose — unless the documented failure fallback above applies (interactive session + the call is unavailable/erroring), in which case the prose fallback is the correct output.
 
 ```
 D<N> — <one-line question title>
@@ -389,6 +389,12 @@ Effort both-scales: when an option involves effort, label both human-team and CC
 
 Net line closes the tradeoff. Per-skill instructions may add stricter rules.
 
+Tool payload rules:
+- `question` is only the decision prompt: one sentence, no newlines, <=80 chars.
+- Background, regrounding, ELI10, stakes, recommendation, pros/cons, and trade-off tables stay in the markdown brief before the tool call.
+- Ask one decision per tool call when possible; batch at most two related questions/tabs. Sequence independent decisions instead of sending 3+ tabs.
+- Do not duplicate the same trade-off text in both `question` and `options[].description`. Prefer putting option-specific trade-offs in `options[].description`.
+
 ### Handling 5+ options — split, never drop
 
 AskUserQuestion caps every call at **4 options**. With 5+ real options, NEVER
@@ -436,6 +442,10 @@ Before calling AskUserQuestion, verify:
 - [ ] (recommended) label on one option (even for neutral-posture)
 - [ ] Dual-scale effort labels on effort-bearing options (human / CC)
 - [ ] Net line closes the decision
+- [ ] `question` is one sentence, no newlines, <=80 chars
+- [ ] Tool call has no more than two related questions/tabs
+- [ ] No duplicated trade-off text between `question` and `options[].description`
+- [ ] You wrote the brief, then called the tool_use payload
 - [ ] You are calling the tool, not writing prose — unless `CONDUCTOR_SESSION: true` (then prose is the DEFAULT, not the tool) OR the documented failure fallback applies (then: prose with the mandatory triad — issue ELI10, per-choice Completeness, Recommendation + `(recommended)` — and a "reply with a letter" instruction, then STOP)
 - [ ] Non-ASCII characters (CJK / accents) written directly, NOT \u-escaped
 - [ ] If you had 5+ options, you split (or batched into ≤4-groups) — did NOT drop any

review/SKILL.md

@@ -341,7 +341,7 @@ Layout: a `D<N>` title + a one-line note to reply with a letter (in Conductor th
 
 ### Format
 
-Every AskUserQuestion is a decision brief and must be sent as tool_use, not prose — unless the documented failure fallback above applies (interactive session + the call is unavailable/erroring), in which case the prose fallback is the correct output.
+Every AskUserQuestion decision has two parts: a markdown decision brief before the call, then a compact tool_use payload. Do not pack the full brief into the tool's `question` string. Every AskUserQuestion is a decision brief and must be sent as tool_use, not prose — unless the documented failure fallback above applies (interactive session + the call is unavailable/erroring), in which case the prose fallback is the correct output.
 
 ```
 D<N> — <one-line question title>
@@ -374,6 +374,12 @@ Effort both-scales: when an option involves effort, label both human-team and CC
 
 Net line closes the tradeoff. Per-skill instructions may add stricter rules.
 
+Tool payload rules:
+- `question` is only the decision prompt: one sentence, no newlines, <=80 chars.
+- Background, regrounding, ELI10, stakes, recommendation, pros/cons, and trade-off tables stay in the markdown brief before the tool call.
+- Ask one decision per tool call when possible; batch at most two related questions/tabs. Sequence independent decisions instead of sending 3+ tabs.
+- Do not duplicate the same trade-off text in both `question` and `options[].description`. Prefer putting option-specific trade-offs in `options[].description`.
+
 ### Handling 5+ options — split, never drop
 
 AskUserQuestion caps every call at **4 options**. With 5+ real options, NEVER
@@ -421,6 +427,10 @@ Before calling AskUserQuestion, verify:
 - [ ] (recommended) label on one option (even for neutral-posture)
 - [ ] Dual-scale effort labels on effort-bearing options (human / CC)
 - [ ] Net line closes the decision
+- [ ] `question` is one sentence, no newlines, <=80 chars
+- [ ] Tool call has no more than two related questions/tabs
+- [ ] No duplicated trade-off text between `question` and `options[].description`
+- [ ] You wrote the brief, then called the tool_use payload
 - [ ] You are calling the tool, not writing prose — unless `CONDUCTOR_SESSION: true` (then prose is the DEFAULT, not the tool) OR the documented failure fallback applies (then: prose with the mandatory triad — issue ELI10, per-choice Completeness, Recommendation + `(recommended)` — and a "reply with a letter" instruction, then STOP)
 - [ ] Non-ASCII characters (CJK / accents) written directly, NOT \u-escaped
 - [ ] If you had 5+ options, you split (or batched into ≤4-groups) — did NOT drop any

scrape/SKILL.md

@@ -337,7 +337,7 @@ Layout: a `D<N>` title + a one-line note to reply with a letter (in Conductor th
 
 ### Format
 
-Every AskUserQuestion is a decision brief and must be sent as tool_use, not prose — unless the documented failure fallback above applies (interactive session + the call is unavailable/erroring), in which case the prose fallback is the correct output.
+Every AskUserQuestion decision has two parts: a markdown decision brief before the call, then a compact tool_use payload. Do not pack the full brief into the tool's `question` string. Every AskUserQuestion is a decision brief and must be sent as tool_use, not prose — unless the documented failure fallback above applies (interactive session + the call is unavailable/erroring), in which case the prose fallback is the correct output.
 
 ```
 D<N> — <one-line question title>
@@ -370,6 +370,12 @@ Effort both-scales: when an option involves effort, label both human-team and CC
 
 Net line closes the tradeoff. Per-skill instructions may add stricter rules.
 
+Tool payload rules:
+- `question` is only the decision prompt: one sentence, no newlines, <=80 chars.
+- Background, regrounding, ELI10, stakes, recommendation, pros/cons, and trade-off tables stay in the markdown brief before the tool call.
+- Ask one decision per tool call when possible; batch at most two related questions/tabs. Sequence independent decisions instead of sending 3+ tabs.
+- Do not duplicate the same trade-off text in both `question` and `options[].description`. Prefer putting option-specific trade-offs in `options[].description`.
+
 ### Handling 5+ options — split, never drop
 
 AskUserQuestion caps every call at **4 options**. With 5+ real options, NEVER
@@ -417,6 +423,10 @@ Before calling AskUserQuestion, verify:
 - [ ] (recommended) label on one option (even for neutral-posture)
 - [ ] Dual-scale effort labels on effort-bearing options (human / CC)
 - [ ] Net line closes the decision
+- [ ] `question` is one sentence, no newlines, <=80 chars
+- [ ] Tool call has no more than two related questions/tabs
+- [ ] No duplicated trade-off text between `question` and `options[].description`
+- [ ] You wrote the brief, then called the tool_use payload
 - [ ] You are calling the tool, not writing prose — unless `CONDUCTOR_SESSION: true` (then prose is the DEFAULT, not the tool) OR the documented failure fallback applies (then: prose with the mandatory triad — issue ELI10, per-choice Completeness, Recommendation + `(recommended)` — and a "reply with a letter" instruction, then STOP)
 - [ ] Non-ASCII characters (CJK / accents) written directly, NOT \u-escaped
 - [ ] If you had 5+ options, you split (or batched into ≤4-groups) — did NOT drop any

scripts/resolvers/preamble/generate-ask-user-format.ts

@@ -39,7 +39,7 @@ Layout: a \`D<N>\` title + a one-line note to reply with a letter (in Conductor
 
 ### Format
 
-Every AskUserQuestion is a decision brief and must be sent as tool_use, not prose — unless the documented failure fallback above applies (interactive session + the call is unavailable/erroring), in which case the prose fallback is the correct output.
+Every AskUserQuestion decision has two parts: a markdown decision brief before the call, then a compact tool_use payload. Do not pack the full brief into the tool's \`question\` string. Every AskUserQuestion is a decision brief and must be sent as tool_use, not prose — unless the documented failure fallback above applies (interactive session + the call is unavailable/erroring), in which case the prose fallback is the correct output.
 
 \`\`\`
 D<N> — <one-line question title>
@@ -72,6 +72,12 @@ Effort both-scales: when an option involves effort, label both human-team and CC
 
 Net line closes the tradeoff. Per-skill instructions may add stricter rules.
 
+Tool payload rules:
+- \`question\` is only the decision prompt: one sentence, no newlines, <=80 chars.
+- Background, regrounding, ELI10, stakes, recommendation, pros/cons, and trade-off tables stay in the markdown brief before the tool call.
+- Ask one decision per tool call when possible; batch at most two related questions/tabs. Sequence independent decisions instead of sending 3+ tabs.
+- Do not duplicate the same trade-off text in both \`question\` and \`options[].description\`. Prefer putting option-specific trade-offs in \`options[].description\`.
+
 ### Handling 5+ options — split, never drop
 
 AskUserQuestion caps every call at **4 options**. With 5+ real options, NEVER
@@ -119,6 +125,10 @@ Before calling AskUserQuestion, verify:
 - [ ] (recommended) label on one option (even for neutral-posture)
 - [ ] Dual-scale effort labels on effort-bearing options (human / CC)
 - [ ] Net line closes the decision
+- [ ] \`question\` is one sentence, no newlines, <=80 chars
+- [ ] Tool call has no more than two related questions/tabs
+- [ ] No duplicated trade-off text between \`question\` and \`options[].description\`
+- [ ] You wrote the brief, then called the tool_use payload
 - [ ] You are calling the tool, not writing prose — unless \`CONDUCTOR_SESSION: true\` (then prose is the DEFAULT, not the tool) OR the documented failure fallback applies (then: prose with the mandatory triad — issue ELI10, per-choice Completeness, Recommendation + \`(recommended)\` — and a "reply with a letter" instruction, then STOP)
 - [ ] Non-ASCII characters (CJK / accents) written directly, NOT \\u-escaped
 - [ ] If you had 5+ options, you split (or batched into ≤4-groups) — did NOT drop any

setup-deploy/SKILL.md

@@ -340,7 +340,7 @@ Layout: a `D<N>` title + a one-line note to reply with a letter (in Conductor th
 
 ### Format
 
-Every AskUserQuestion is a decision brief and must be sent as tool_use, not prose — unless the documented failure fallback above applies (interactive session + the call is unavailable/erroring), in which case the prose fallback is the correct output.
+Every AskUserQuestion decision has two parts: a markdown decision brief before the call, then a compact tool_use payload. Do not pack the full brief into the tool's `question` string. Every AskUserQuestion is a decision brief and must be sent as tool_use, not prose — unless the documented failure fallback above applies (interactive session + the call is unavailable/erroring), in which case the prose fallback is the correct output.
 
 ```
 D<N> — <one-line question title>
@@ -373,6 +373,12 @@ Effort both-scales: when an option involves effort, label both human-team and CC
 
 Net line closes the tradeoff. Per-skill instructions may add stricter rules.
 
+Tool payload rules:
+- `question` is only the decision prompt: one sentence, no newlines, <=80 chars.
+- Background, regrounding, ELI10, stakes, recommendation, pros/cons, and trade-off tables stay in the markdown brief before the tool call.
+- Ask one decision per tool call when possible; batch at most two related questions/tabs. Sequence independent decisions instead of sending 3+ tabs.
+- Do not duplicate the same trade-off text in both `question` and `options[].description`. Prefer putting option-specific trade-offs in `options[].description`.
+
 ### Handling 5+ options — split, never drop
 
 AskUserQuestion caps every call at **4 options**. With 5+ real options, NEVER
@@ -420,6 +426,10 @@ Before calling AskUserQuestion, verify:
 - [ ] (recommended) label on one option (even for neutral-posture)
 - [ ] Dual-scale effort labels on effort-bearing options (human / CC)
 - [ ] Net line closes the decision
+- [ ] `question` is one sentence, no newlines, <=80 chars
+- [ ] Tool call has no more than two related questions/tabs
+- [ ] No duplicated trade-off text between `question` and `options[].description`
+- [ ] You wrote the brief, then called the tool_use payload
 - [ ] You are calling the tool, not writing prose — unless `CONDUCTOR_SESSION: true` (then prose is the DEFAULT, not the tool) OR the documented failure fallback applies (then: prose with the mandatory triad — issue ELI10, per-choice Completeness, Recommendation + `(recommended)` — and a "reply with a letter" instruction, then STOP)
 - [ ] Non-ASCII characters (CJK / accents) written directly, NOT \u-escaped
 - [ ] If you had 5+ options, you split (or batched into ≤4-groups) — did NOT drop any

setup-gbrain/SKILL.md

@@ -339,7 +339,7 @@ Layout: a `D<N>` title + a one-line note to reply with a letter (in Conductor th
 
 ### Format
 
-Every AskUserQuestion is a decision brief and must be sent as tool_use, not prose — unless the documented failure fallback above applies (interactive session + the call is unavailable/erroring), in which case the prose fallback is the correct output.
+Every AskUserQuestion decision has two parts: a markdown decision brief before the call, then a compact tool_use payload. Do not pack the full brief into the tool's `question` string. Every AskUserQuestion is a decision brief and must be sent as tool_use, not prose — unless the documented failure fallback above applies (interactive session + the call is unavailable/erroring), in which case the prose fallback is the correct output.
 
 ```
 D<N> — <one-line question title>
@@ -372,6 +372,12 @@ Effort both-scales: when an option involves effort, label both human-team and CC
 
 Net line closes the tradeoff. Per-skill instructions may add stricter rules.
 
+Tool payload rules:
+- `question` is only the decision prompt: one sentence, no newlines, <=80 chars.
+- Background, regrounding, ELI10, stakes, recommendation, pros/cons, and trade-off tables stay in the markdown brief before the tool call.
+- Ask one decision per tool call when possible; batch at most two related questions/tabs. Sequence independent decisions instead of sending 3+ tabs.
+- Do not duplicate the same trade-off text in both `question` and `options[].description`. Prefer putting option-specific trade-offs in `options[].description`.
+
 ### Handling 5+ options — split, never drop
 
 AskUserQuestion caps every call at **4 options**. With 5+ real options, NEVER
@@ -419,6 +425,10 @@ Before calling AskUserQuestion, verify:
 - [ ] (recommended) label on one option (even for neutral-posture)
 - [ ] Dual-scale effort labels on effort-bearing options (human / CC)
 - [ ] Net line closes the decision
+- [ ] `question` is one sentence, no newlines, <=80 chars
+- [ ] Tool call has no more than two related questions/tabs
+- [ ] No duplicated trade-off text between `question` and `options[].description`
+- [ ] You wrote the brief, then called the tool_use payload
 - [ ] You are calling the tool, not writing prose — unless `CONDUCTOR_SESSION: true` (then prose is the DEFAULT, not the tool) OR the documented failure fallback applies (then: prose with the mandatory triad — issue ELI10, per-choice Completeness, Recommendation + `(recommended)` — and a "reply with a letter" instruction, then STOP)
 - [ ] Non-ASCII characters (CJK / accents) written directly, NOT \u-escaped
 - [ ] If you had 5+ options, you split (or batched into ≤4-groups) — did NOT drop any

ship/SKILL.md

@@ -341,7 +341,7 @@ Layout: a `D<N>` title + a one-line note to reply with a letter (in Conductor th
 
 ### Format
 
-Every AskUserQuestion is a decision brief and must be sent as tool_use, not prose — unless the documented failure fallback above applies (interactive session + the call is unavailable/erroring), in which case the prose fallback is the correct output.
+Every AskUserQuestion decision has two parts: a markdown decision brief before the call, then a compact tool_use payload. Do not pack the full brief into the tool's `question` string. Every AskUserQuestion is a decision brief and must be sent as tool_use, not prose — unless the documented failure fallback above applies (interactive session + the call is unavailable/erroring), in which case the prose fallback is the correct output.
 
 ```
 D<N> — <one-line question title>
@@ -374,6 +374,12 @@ Effort both-scales: when an option involves effort, label both human-team and CC
 
 Net line closes the tradeoff. Per-skill instructions may add stricter rules.
 
+Tool payload rules:
+- `question` is only the decision prompt: one sentence, no newlines, <=80 chars.
+- Background, regrounding, ELI10, stakes, recommendation, pros/cons, and trade-off tables stay in the markdown brief before the tool call.
+- Ask one decision per tool call when possible; batch at most two related questions/tabs. Sequence independent decisions instead of sending 3+ tabs.
+- Do not duplicate the same trade-off text in both `question` and `options[].description`. Prefer putting option-specific trade-offs in `options[].description`.
+
 ### Handling 5+ options — split, never drop
 
 AskUserQuestion caps every call at **4 options**. With 5+ real options, NEVER
@@ -421,6 +427,10 @@ Before calling AskUserQuestion, verify:
 - [ ] (recommended) label on one option (even for neutral-posture)
 - [ ] Dual-scale effort labels on effort-bearing options (human / CC)
 - [ ] Net line closes the decision
+- [ ] `question` is one sentence, no newlines, <=80 chars
+- [ ] Tool call has no more than two related questions/tabs
+- [ ] No duplicated trade-off text between `question` and `options[].description`
+- [ ] You wrote the brief, then called the tool_use payload
 - [ ] You are calling the tool, not writing prose — unless `CONDUCTOR_SESSION: true` (then prose is the DEFAULT, not the tool) OR the documented failure fallback applies (then: prose with the mandatory triad — issue ELI10, per-choice Completeness, Recommendation + `(recommended)` — and a "reply with a letter" instruction, then STOP)
 - [ ] Non-ASCII characters (CJK / accents) written directly, NOT \u-escaped
 - [ ] If you had 5+ options, you split (or batched into ≤4-groups) — did NOT drop any

skillify/SKILL.md

@@ -337,7 +337,7 @@ Layout: a `D<N>` title + a one-line note to reply with a letter (in Conductor th
 
 ### Format
 
-Every AskUserQuestion is a decision brief and must be sent as tool_use, not prose — unless the documented failure fallback above applies (interactive session + the call is unavailable/erroring), in which case the prose fallback is the correct output.
+Every AskUserQuestion decision has two parts: a markdown decision brief before the call, then a compact tool_use payload. Do not pack the full brief into the tool's `question` string. Every AskUserQuestion is a decision brief and must be sent as tool_use, not prose — unless the documented failure fallback above applies (interactive session + the call is unavailable/erroring), in which case the prose fallback is the correct output.
 
 ```
 D<N> — <one-line question title>
@@ -370,6 +370,12 @@ Effort both-scales: when an option involves effort, label both human-team and CC
 
 Net line closes the tradeoff. Per-skill instructions may add stricter rules.
 
+Tool payload rules:
+- `question` is only the decision prompt: one sentence, no newlines, <=80 chars.
+- Background, regrounding, ELI10, stakes, recommendation, pros/cons, and trade-off tables stay in the markdown brief before the tool call.
+- Ask one decision per tool call when possible; batch at most two related questions/tabs. Sequence independent decisions instead of sending 3+ tabs.
+- Do not duplicate the same trade-off text in both `question` and `options[].description`. Prefer putting option-specific trade-offs in `options[].description`.
+
 ### Handling 5+ options — split, never drop
 
 AskUserQuestion caps every call at **4 options**. With 5+ real options, NEVER
@@ -417,6 +423,10 @@ Before calling AskUserQuestion, verify:
 - [ ] (recommended) label on one option (even for neutral-posture)
 - [ ] Dual-scale effort labels on effort-bearing options (human / CC)
 - [ ] Net line closes the decision
+- [ ] `question` is one sentence, no newlines, <=80 chars
+- [ ] Tool call has no more than two related questions/tabs
+- [ ] No duplicated trade-off text between `question` and `options[].description`
+- [ ] You wrote the brief, then called the tool_use payload
 - [ ] You are calling the tool, not writing prose — unless `CONDUCTOR_SESSION: true` (then prose is the DEFAULT, not the tool) OR the documented failure fallback applies (then: prose with the mandatory triad — issue ELI10, per-choice Completeness, Recommendation + `(recommended)` — and a "reply with a letter" instruction, then STOP)
 - [ ] Non-ASCII characters (CJK / accents) written directly, NOT \u-escaped
 - [ ] If you had 5+ options, you split (or batched into ≤4-groups) — did NOT drop any

spec/SKILL.md

@@ -338,7 +338,7 @@ Layout: a `D<N>` title + a one-line note to reply with a letter (in Conductor th
 
 ### Format
 
-Every AskUserQuestion is a decision brief and must be sent as tool_use, not prose — unless the documented failure fallback above applies (interactive session + the call is unavailable/erroring), in which case the prose fallback is the correct output.
+Every AskUserQuestion decision has two parts: a markdown decision brief before the call, then a compact tool_use payload. Do not pack the full brief into the tool's `question` string. Every AskUserQuestion is a decision brief and must be sent as tool_use, not prose — unless the documented failure fallback above applies (interactive session + the call is unavailable/erroring), in which case the prose fallback is the correct output.
 
 ```
 D<N> — <one-line question title>
@@ -371,6 +371,12 @@ Effort both-scales: when an option involves effort, label both human-team and CC
 
 Net line closes the tradeoff. Per-skill instructions may add stricter rules.
 
+Tool payload rules:
+- `question` is only the decision prompt: one sentence, no newlines, <=80 chars.
+- Background, regrounding, ELI10, stakes, recommendation, pros/cons, and trade-off tables stay in the markdown brief before the tool call.
+- Ask one decision per tool call when possible; batch at most two related questions/tabs. Sequence independent decisions instead of sending 3+ tabs.
+- Do not duplicate the same trade-off text in both `question` and `options[].description`. Prefer putting option-specific trade-offs in `options[].description`.
+
 ### Handling 5+ options — split, never drop
 
 AskUserQuestion caps every call at **4 options**. With 5+ real options, NEVER
@@ -418,6 +424,10 @@ Before calling AskUserQuestion, verify:
 - [ ] (recommended) label on one option (even for neutral-posture)
 - [ ] Dual-scale effort labels on effort-bearing options (human / CC)
 - [ ] Net line closes the decision
+- [ ] `question` is one sentence, no newlines, <=80 chars
+- [ ] Tool call has no more than two related questions/tabs
+- [ ] No duplicated trade-off text between `question` and `options[].description`
+- [ ] You wrote the brief, then called the tool_use payload
 - [ ] You are calling the tool, not writing prose — unless `CONDUCTOR_SESSION: true` (then prose is the DEFAULT, not the tool) OR the documented failure fallback applies (then: prose with the mandatory triad — issue ELI10, per-choice Completeness, Recommendation + `(recommended)` — and a "reply with a letter" instruction, then STOP)
 - [ ] Non-ASCII characters (CJK / accents) written directly, NOT \u-escaped
 - [ ] If you had 5+ options, you split (or batched into ≤4-groups) — did NOT drop any
@@ -1379,7 +1389,7 @@ Layout: a `D<N>` title + a one-line note to reply with a letter (in Conductor th
 
 ### Format
 
-Every AskUserQuestion is a decision brief and must be sent as tool_use, not prose — unless the documented failure fallback above applies (interactive session + the call is unavailable/erroring), in which case the prose fallback is the correct output.
+Every AskUserQuestion decision has two parts: a markdown decision brief before the call, then a compact tool_use payload. Do not pack the full brief into the tool's `question` string. Every AskUserQuestion is a decision brief and must be sent as tool_use, not prose — unless the documented failure fallback above applies (interactive session + the call is unavailable/erroring), in which case the prose fallback is the correct output.
 
 ```
 D<N> — <one-line question title>
@@ -1412,6 +1422,12 @@ Effort both-scales: when an option involves effort, label both human-team and CC
 
 Net line closes the tradeoff. Per-skill instructions may add stricter rules.
 
+Tool payload rules:
+- `question` is only the decision prompt: one sentence, no newlines, <=80 chars.
+- Background, regrounding, ELI10, stakes, recommendation, pros/cons, and trade-off tables stay in the markdown brief before the tool call.
+- Ask one decision per tool call when possible; batch at most two related questions/tabs. Sequence independent decisions instead of sending 3+ tabs.
+- Do not duplicate the same trade-off text in both `question` and `options[].description`. Prefer putting option-specific trade-offs in `options[].description`.
+
 ### Handling 5+ options — split, never drop
 
 AskUserQuestion caps every call at **4 options**. With 5+ real options, NEVER
@@ -1459,6 +1475,10 @@ Before calling AskUserQuestion, verify:
 - [ ] (recommended) label on one option (even for neutral-posture)
 - [ ] Dual-scale effort labels on effort-bearing options (human / CC)
 - [ ] Net line closes the decision
+- [ ] `question` is one sentence, no newlines, <=80 chars
+- [ ] Tool call has no more than two related questions/tabs
+- [ ] No duplicated trade-off text between `question` and `options[].description`
+- [ ] You wrote the brief, then called the tool_use payload
 - [ ] You are calling the tool, not writing prose — unless `CONDUCTOR_SESSION: true` (then prose is the DEFAULT, not the tool) OR the documented failure fallback applies (then: prose with the mandatory triad — issue ELI10, per-choice Completeness, Recommendation + `(recommended)` — and a "reply with a letter" instruction, then STOP)
 - [ ] Non-ASCII characters (CJK / accents) written directly, NOT \u-escaped
 - [ ] If you had 5+ options, you split (or batched into ≤4-groups) — did NOT drop any

sync-gbrain/SKILL.md

@@ -339,7 +339,7 @@ Layout: a `D<N>` title + a one-line note to reply with a letter (in Conductor th
 
 ### Format
 
-Every AskUserQuestion is a decision brief and must be sent as tool_use, not prose — unless the documented failure fallback above applies (interactive session + the call is unavailable/erroring), in which case the prose fallback is the correct output.
+Every AskUserQuestion decision has two parts: a markdown decision brief before the call, then a compact tool_use payload. Do not pack the full brief into the tool's `question` string. Every AskUserQuestion is a decision brief and must be sent as tool_use, not prose — unless the documented failure fallback above applies (interactive session + the call is unavailable/erroring), in which case the prose fallback is the correct output.
 
 ```
 D<N> — <one-line question title>
@@ -372,6 +372,12 @@ Effort both-scales: when an option involves effort, label both human-team and CC
 
 Net line closes the tradeoff. Per-skill instructions may add stricter rules.
 
+Tool payload rules:
+- `question` is only the decision prompt: one sentence, no newlines, <=80 chars.
+- Background, regrounding, ELI10, stakes, recommendation, pros/cons, and trade-off tables stay in the markdown brief before the tool call.
+- Ask one decision per tool call when possible; batch at most two related questions/tabs. Sequence independent decisions instead of sending 3+ tabs.
+- Do not duplicate the same trade-off text in both `question` and `options[].description`. Prefer putting option-specific trade-offs in `options[].description`.
+
 ### Handling 5+ options — split, never drop
 
 AskUserQuestion caps every call at **4 options**. With 5+ real options, NEVER
@@ -419,6 +425,10 @@ Before calling AskUserQuestion, verify:
 - [ ] (recommended) label on one option (even for neutral-posture)
 - [ ] Dual-scale effort labels on effort-bearing options (human / CC)
 - [ ] Net line closes the decision
+- [ ] `question` is one sentence, no newlines, <=80 chars
+- [ ] Tool call has no more than two related questions/tabs
+- [ ] No duplicated trade-off text between `question` and `options[].description`
+- [ ] You wrote the brief, then called the tool_use payload
 - [ ] You are calling the tool, not writing prose — unless `CONDUCTOR_SESSION: true` (then prose is the DEFAULT, not the tool) OR the documented failure fallback applies (then: prose with the mandatory triad — issue ELI10, per-choice Completeness, Recommendation + `(recommended)` — and a "reply with a letter" instruction, then STOP)
 - [ ] Non-ASCII characters (CJK / accents) written directly, NOT \u-escaped
 - [ ] If you had 5+ options, you split (or batched into ≤4-groups) — did NOT drop any

test/fixtures/golden/claude-ship-SKILL.md

@@ -341,7 +341,7 @@ Layout: a `D<N>` title + a one-line note to reply with a letter (in Conductor th
 
 ### Format
 
-Every AskUserQuestion is a decision brief and must be sent as tool_use, not prose — unless the documented failure fallback above applies (interactive session + the call is unavailable/erroring), in which case the prose fallback is the correct output.
+Every AskUserQuestion decision has two parts: a markdown decision brief before the call, then a compact tool_use payload. Do not pack the full brief into the tool's `question` string. Every AskUserQuestion is a decision brief and must be sent as tool_use, not prose — unless the documented failure fallback above applies (interactive session + the call is unavailable/erroring), in which case the prose fallback is the correct output.
 
 ```
 D<N> — <one-line question title>
@@ -374,6 +374,12 @@ Effort both-scales: when an option involves effort, label both human-team and CC
 
 Net line closes the tradeoff. Per-skill instructions may add stricter rules.
 
+Tool payload rules:
+- `question` is only the decision prompt: one sentence, no newlines, <=80 chars.
+- Background, regrounding, ELI10, stakes, recommendation, pros/cons, and trade-off tables stay in the markdown brief before the tool call.
+- Ask one decision per tool call when possible; batch at most two related questions/tabs. Sequence independent decisions instead of sending 3+ tabs.
+- Do not duplicate the same trade-off text in both `question` and `options[].description`. Prefer putting option-specific trade-offs in `options[].description`.
+
 ### Handling 5+ options — split, never drop
 
 AskUserQuestion caps every call at **4 options**. With 5+ real options, NEVER
@@ -421,6 +427,10 @@ Before calling AskUserQuestion, verify:
 - [ ] (recommended) label on one option (even for neutral-posture)
 - [ ] Dual-scale effort labels on effort-bearing options (human / CC)
 - [ ] Net line closes the decision
+- [ ] `question` is one sentence, no newlines, <=80 chars
+- [ ] Tool call has no more than two related questions/tabs
+- [ ] No duplicated trade-off text between `question` and `options[].description`
+- [ ] You wrote the brief, then called the tool_use payload
 - [ ] You are calling the tool, not writing prose — unless `CONDUCTOR_SESSION: true` (then prose is the DEFAULT, not the tool) OR the documented failure fallback applies (then: prose with the mandatory triad — issue ELI10, per-choice Completeness, Recommendation + `(recommended)` — and a "reply with a letter" instruction, then STOP)
 - [ ] Non-ASCII characters (CJK / accents) written directly, NOT \u-escaped
 - [ ] If you had 5+ options, you split (or batched into ≤4-groups) — did NOT drop any

test/fixtures/golden/codex-ship-SKILL.md

@@ -327,7 +327,7 @@ Layout: a `D<N>` title + a one-line note to reply with a letter (in Conductor th
 
 ### Format
 
-Every AskUserQuestion is a decision brief and must be sent as tool_use, not prose — unless the documented failure fallback above applies (interactive session + the call is unavailable/erroring), in which case the prose fallback is the correct output.
+Every AskUserQuestion decision has two parts: a markdown decision brief before the call, then a compact tool_use payload. Do not pack the full brief into the tool's `question` string. Every AskUserQuestion is a decision brief and must be sent as tool_use, not prose — unless the documented failure fallback above applies (interactive session + the call is unavailable/erroring), in which case the prose fallback is the correct output.
 
 ```
 D<N> — <one-line question title>
@@ -360,6 +360,12 @@ Effort both-scales: when an option involves effort, label both human-team and CC
 
 Net line closes the tradeoff. Per-skill instructions may add stricter rules.
 
+Tool payload rules:
+- `question` is only the decision prompt: one sentence, no newlines, <=80 chars.
+- Background, regrounding, ELI10, stakes, recommendation, pros/cons, and trade-off tables stay in the markdown brief before the tool call.
+- Ask one decision per tool call when possible; batch at most two related questions/tabs. Sequence independent decisions instead of sending 3+ tabs.
+- Do not duplicate the same trade-off text in both `question` and `options[].description`. Prefer putting option-specific trade-offs in `options[].description`.
+
 ### Handling 5+ options — split, never drop
 
 AskUserQuestion caps every call at **4 options**. With 5+ real options, NEVER
@@ -407,6 +413,10 @@ Before calling AskUserQuestion, verify:
 - [ ] (recommended) label on one option (even for neutral-posture)
 - [ ] Dual-scale effort labels on effort-bearing options (human / CC)
 - [ ] Net line closes the decision
+- [ ] `question` is one sentence, no newlines, <=80 chars
+- [ ] Tool call has no more than two related questions/tabs
+- [ ] No duplicated trade-off text between `question` and `options[].description`
+- [ ] You wrote the brief, then called the tool_use payload
 - [ ] You are calling the tool, not writing prose — unless `CONDUCTOR_SESSION: true` (then prose is the DEFAULT, not the tool) OR the documented failure fallback applies (then: prose with the mandatory triad — issue ELI10, per-choice Completeness, Recommendation + `(recommended)` — and a "reply with a letter" instruction, then STOP)
 - [ ] Non-ASCII characters (CJK / accents) written directly, NOT \u-escaped
 - [ ] If you had 5+ options, you split (or batched into ≤4-groups) — did NOT drop any

test/fixtures/golden/factory-ship-SKILL.md

@@ -329,7 +329,7 @@ Layout: a `D<N>` title + a one-line note to reply with a letter (in Conductor th
 
 ### Format
 
-Every AskUserQuestion is a decision brief and must be sent as tool_use, not prose — unless the documented failure fallback above applies (interactive session + the call is unavailable/erroring), in which case the prose fallback is the correct output.
+Every AskUserQuestion decision has two parts: a markdown decision brief before the call, then a compact tool_use payload. Do not pack the full brief into the tool's `question` string. Every AskUserQuestion is a decision brief and must be sent as tool_use, not prose — unless the documented failure fallback above applies (interactive session + the call is unavailable/erroring), in which case the prose fallback is the correct output.
 
 ```
 D<N> — <one-line question title>
@@ -362,6 +362,12 @@ Effort both-scales: when an option involves effort, label both human-team and CC
 
 Net line closes the tradeoff. Per-skill instructions may add stricter rules.
 
+Tool payload rules:
+- `question` is only the decision prompt: one sentence, no newlines, <=80 chars.
+- Background, regrounding, ELI10, stakes, recommendation, pros/cons, and trade-off tables stay in the markdown brief before the tool call.
+- Ask one decision per tool call when possible; batch at most two related questions/tabs. Sequence independent decisions instead of sending 3+ tabs.
+- Do not duplicate the same trade-off text in both `question` and `options[].description`. Prefer putting option-specific trade-offs in `options[].description`.
+
 ### Handling 5+ options — split, never drop
 
 AskUserQuestion caps every call at **4 options**. With 5+ real options, NEVER
@@ -409,6 +415,10 @@ Before calling AskUserQuestion, verify:
 - [ ] (recommended) label on one option (even for neutral-posture)
 - [ ] Dual-scale effort labels on effort-bearing options (human / CC)
 - [ ] Net line closes the decision
+- [ ] `question` is one sentence, no newlines, <=80 chars
+- [ ] Tool call has no more than two related questions/tabs
+- [ ] No duplicated trade-off text between `question` and `options[].description`
+- [ ] You wrote the brief, then called the tool_use payload
 - [ ] You are calling the tool, not writing prose — unless `CONDUCTOR_SESSION: true` (then prose is the DEFAULT, not the tool) OR the documented failure fallback applies (then: prose with the mandatory triad — issue ELI10, per-choice Completeness, Recommendation + `(recommended)` — and a "reply with a letter" instruction, then STOP)
 - [ ] Non-ASCII characters (CJK / accents) written directly, NOT \u-escaped
 - [ ] If you had 5+ options, you split (or batched into ≤4-groups) — did NOT drop any

test/gen-skill-docs.test.ts

@@ -332,6 +332,17 @@ describe('gen-skill-docs', () => {
     expect(content).toContain('not function names');
   });
 
+  // #1208: the compact-payload rules must reach generated tier 2+ SKILL.md so
+  // panel hosts (VSCode) render readable question cards.
+  test('tier 2+ skills keep AskUserQuestion panel payloads compact', () => {
+    const content = fs.readFileSync(path.join(ROOT, 'cso', 'SKILL.md'), 'utf-8');
+    expect(content).toContain("Do not pack the full brief into the tool's `question` string");
+    expect(content).toContain('`question` is only the decision prompt');
+    expect(content).toContain('<=80 chars');
+    expect(content).toContain('batch at most two related questions/tabs');
+    expect(content).toContain('No duplicated trade-off text');
+  });
+
   test('tier 1 skills do NOT contain AskUserQuestion format', () => {
     // Use benchmark (tier 1) instead of root — root SKILL.md gets overwritten by Codex test setup
     const content = fs.readFileSync(path.join(ROOT, 'benchmark', 'SKILL.md'), 'utf-8');

test/resolver-ask-user-format.test.ts

@@ -105,10 +105,35 @@ describe('generateAskUserFormat — v1.7.0.0 Pros/Cons format', () => {
     expect(out).toMatch(/not a[\s\S]*question|not[\s\S]*interactive/i);
   });
 
+  // #1208: keep the tool_use payload compact so the question card stays
+  // readable in panel hosts (VSCode). The full brief belongs in markdown
+  // before the call, not inside the tool's `question` string.
+  test('keeps long decision briefs out of the tool question field', () => {
+    expect(out).toMatch(/Do not pack the full brief into the tool's `question` string/);
+    expect(out).toMatch(/`question` is only the decision prompt/);
+    expect(out).toMatch(/<=80 chars/);
+    expect(out).toMatch(/no newlines/);
+  });
+
+  test('limits batched AskUserQuestion tabs for panel readability', () => {
+    expect(out).toMatch(/batch at most two related questions\/tabs/i);
+    expect(out).toMatch(/Sequence independent decisions/i);
+  });
+
+  test('forbids duplicate trade-off text in question and option descriptions', () => {
+    expect(out).toMatch(/Do not duplicate the same trade-off text/);
+    expect(out).toMatch(/options\[\]\.description/);
+  });
+
   test('includes self-check before emitting', () => {
     expect(out).toContain('Self-check before emitting');
     expect(out).toMatch(/D<N> header present/);
     expect(out).toMatch(/Net line closes/);
+    // #1208 self-check items for compact payloads.
+    expect(out).toMatch(/`question` is one sentence/);
+    expect(out).toMatch(/no more than two related questions\/tabs/);
+    expect(out).toMatch(/No duplicated trade-off text/);
+    expect(out).toMatch(/wrote the brief, then called the tool_use payload/);
   });
 
   test('documents D-numbering as model-level not runtime state', () => {