Inspired by Claude Code v2.1.223: 'Fixed permission prompts so commands
padded with tabs or invisible Unicode can no longer hide part of the
command from the approval dialog.'
A dangerous command rendered into an approval prompt could previously
lie to the human approver three ways:
- invisible/format Unicode (zero-width, bidi overrides/isolates,
variation selectors, U+E0000 tag block) rendered as nothing
- raw control bytes (ANSI/OSC escapes, bare CR) could erase or
overwrite the just-printed prompt line in the terminal
- long whitespace padding runs pushed the dangerous tail out of view
or past platform preview truncation (~200 chars on gateway)
New agent.redact.sanitize_command_for_display() replaces hidden chars
with visible escape markers (\u202e, \x1b) and collapses padding runs
to explicit markers, preserving literal IOCs instead of deleting them.
Wired at every approval display-mint site: CLI prompt, gateway
dangerous-command + execute_code + tool-approval payloads, pending
fallbacks, and gateway _redact_approval_command. Display-only — the
executed command and pattern-key persistence are untouched.
25 new tests; redact (97), approval (103), gateway approval-format
suites green; E2E with real imports across CLI + gateway paths.
Phase 8 of the Hermes Agent token leak mitigation plan
(ralplan-hermes-token-leaks.md §3.9). Adds a boolean kwarg
`skip_background_review` (default False) to AIAgent.__init__ that
suppresses the end-of-turn _spawn_background_review fork.
Each background review fork instantiates a new AIAgent with its own
~15K input tokens + up to 8 LLM iterations, accumulating ~30K tokens
per event in the worst case. On cron sessions there is no
human-in-the-loop benefit from the review (no skill-creation pressure,
nobody curating MEMORY.md), so the cost is pure waste.
The end-of-turn guard now reads:
if (final_response and not interrupted
and not getattr(self, "skip_background_review", False)
and (_should_review_memory or _should_review_skills)):
skip_memory=True already disables the memory-review trigger; this
flag is the explicit single-switch off for both review paths.
Defaults to False, so behavior is unchanged for gateway/CLI callers
that omit the kwarg.
Tests: 5 new unit tests in tests/agent/test_skip_background_review.py
covering the default value, flag persistence, the gate short-circuit,
the gate fall-through, and a source-text assertion that the cron
scheduler sets the flag to True (separate commit).
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
Rescope: hermes verify fills only the runtime-smoke gap and plugs into
the pieces Hermes already has instead of standing beside them.
- agent/verification_evidence.py: record_verify_run() — explicit ledger
write for hermes verify results (shared _insert_evidence factored out
of record_terminal_result). Passing runs mark the workspace passed
like scripts/run_tests.sh; failures are recorded; --phase/--skip-start
runs are recorded as targeted scope.
- hermes_cli/verify_cmd.py: record results into the ledger on completion
(fail-silent, HERMES_SESSION_ID attribution); on the detect path merge
detect_project_facts verify commands the recipe missed into the
recipe's test list (never applied to a saved manifest).
- agent/verification_stop.py: recipe-aware nudge — when the workspace
has a runnable recipe (start command or .hermes/environment.json),
suggest hermes verify --json as the preferred full check; cheap,
try/except-guarded detection that can never break the nudge path.
- agent/verify/recipes.py: document layer ownership (coding_context =
cheap prompt facts; verify/recipes = deep runtime recipe).
- tests/verify/test_ledger_and_nudge_integration.py: 17 tests covering
ledger pass/fail recording, the closed edit->nudge->verify->satisfied
loop, recipe-aware nudge wording + fail-silence, and the facts merge.
Resumable ESTOP sentinel at $HERMES_HOME/ESTOP that halts NEW work only:
- agent/estop.py: sentinel engage/disengage/is_engaged (single stat, no
caching), optional reason + timestamp stored as JSON, paused_reply()
notice, check_paused() log-once-per-engagement helper. Corrupt/empty
sentinel still pauses (fail safe); a `touch ~/.hermes/ESTOP` works.
- cron/scheduler.py: tick() skips dispatch while engaged (logged once per
engagement, not per tick). Due jobs simply wait for the next tick after
resume — in-flight runs are never touched.
- gateway/kanban_watchers.py: dispatcher skips auto-decompose and worker
spawning while engaged; zombie reaping still runs and running workers
finish naturally.
- gateway/run.py: new gateway turns (post-auth, non-internal) get a brief
"Hermes is paused" reply instead of an agent run. Internal events
(in-flight background completions) bypass the gate.
- hermes_cli/subcommands/pause.py: `hermes pause [--reason]` and
`hermes resume`, wired into main() and _BUILTIN_SUBCOMMANDS.
- hermes_cli/status.py: `hermes status` shows a PAUSED banner (one stat).
- tests/test_estop.py: 20 tests — sentinel lifecycle, reason surfacing,
log-once, cron skip + resume, kanban gate, gateway paused reply +
internal bypass, CLI idempotence, builtin-set parity, status line.
Never kills in-flight work; resumable with no restart. Footprint ladder:
CLI command only, no new model tool, no new env vars.
Ported from: gastownhall/gastown estop.go (MIT); related prior art:
#26778 (/panic — kill/exit semantics, deliberately different: ours is a
resumable pause), #44617 (interrupt in-flight cron — out of scope here).
When the Nous Portal returns paid_service_access.allowed=false with
reason=member_spend_cap_exceeded, Hermes was falling through to the
generic 'no active subscription or usable credits' message — even
though the user has ample purchased credits and the real blocker is
an org-level per-member spend cap.
This adds a dedicated branch that surfaces the actual cause: names the
spend cap, shows the cap/spend amounts, and tells the user to ask their
org admin to raise it. Also adds member_spend_cap_exceeded to the
billing error code set so the error classifier and auth error formatter
route it through the Nous entitlement message path.
Follow-up to the review on the session.resume ownership fix. Closing the
pre-transfer early returns left two gaps, both real.
1. The transfer had no owner on the other side. Once ownership moved to the
agent, teardown ran AIAgent.close() (via _teardown_session on session.close
and the orphaned-session reaper), which called session_db.end_session() —
that finalizes the session ROW, not the connection. A successfully resumed
profile session kept its dedicated handle, its db/-wal/-shm fds and its
background token-writer thread for the life of the gateway.
AIAgent now carries an explicit _owns_session_db, defaulting False so the
SHARED launch handle — which outlives every agent and backs every other live
session — is still never closed there. Only the dedicated-open sites set it,
at the point ownership actually changes hands.
2. session.resume was not the only profile-scoped open with no close on its
failure paths. Covered here with the same flag, via a _transfer_db_to_agent
helper that refuses the transfer unless the agent really holds that handle:
- the deferred builder (_start_agent_build), including the session-reaped-
mid-build case, where the built agent is discarded and never torn down, so
transferring to it would leak exactly as before;
- session.branch's branch_db;
- the compute host's per-profile open;
- AIAgent's own lazy open in _get_session_db_for_recall, which no other
object ever references and so was unconditionally abandoned.
Where a handle has already reached a registered session, the drop is
unconditional and the transfer is best-effort on top: a refused transfer leaves
the old leak, which is survivable, whereas closing under a live session is the
permanent "Cannot operate on a closed database" break the original patch exists
to avoid.
Tests: tests/tui_gateway/test_session_db_ownership_teardown.py (new, 14).
11 of the 14 fail without this change; the 3 that pass are the "must NOT close"
guards, which hold in both directions by design.
Review-pass follow-ups (three parallel reviewers, findings verified):
- hermes_state_search.py list_recent_user_messages now drops legacy
standalone compaction handoffs in the decode loop (SQL can't see them:
durable role=user, no display_kind). Closes the /undo N pairing skew
where the in-memory count (new predicate) and the DB soft-delete pick
(old predicate) targeted different turns on legacy sessions. Fetches
with headroom so the requested limit is still honored. 3 new tests,
mutation-checked (no-op'ing the skip fails 2/3).
- _should_skip_model_call_for_reference_handoff: single drive-check scan
(was two — once inside the restore helper, once after); the restore
helper no longer re-scans and its return value now decides the verdict.
- _final_response_from_messages replaced by the _HANDOFF_SKIP_FINAL_RESPONSE
constant it always returned (parameter was unused).
- _handoff_carries_live_user_content delegates to the canonical
_strip_context_summary_handoff_message — also fixes the edge where a
merged-shaped row with an EMPTY preserved prior tail was wrongly
treated as carrying live content.
- Site-level guard test for rollback.restore with a legacy handoff row
(predicate-in-context, complements the unit tests).
Follow-ups on top of the salvaged #80696 fix (review findings):
- Sibling sites: rollback.restore, gateway /retry, CLI /retry and /undo N,
and both CLI resume turn counters now use is_user_originated_turn so
legacy-persisted standalone handoffs (durable role=user, no display_kind)
can never be truncation targets or counted as user turns (#80622
suggested regression 4, dispatcher-wide).
- Site-1 guard: hoist the api_call_count decrement + iteration-budget
refund above the break so a skipped turn no longer leaks a budget unit
and finalize_turn logs the true call count (matches the ollama early-exit
and the site-2 sibling).
- Site-2 guard: run the handoff guard BEFORE reanchoring so a restored
user ask is what the anchor lands on, not a stale pre-restore index.
- SUMMARY_PREFIX: add the mid-tool-loop carve-out the code-side guard
already implements, so a literal-minded model doesn't halt an in-flight
exchange after in-place compaction.
- Skip path returns a short compaction status instead of replaying the
previous turn's answer (finalize_turn would append it as a fresh
assistant row — duplicate prose in transcript and delivery).
After a completed assistant stop, a standalone CONTEXT COMPACTION handoff
could occupy the sole user slot and resume stale Historical Task Snapshot
work with no new human ask. Guard post-compaction continues, hide
standalone handoffs from session dispatch, and harden SUMMARY_PREFIX for
the empty-after-handoff case (#80622).
Two docstring corrections on top of the salvaged #80997 fix — behavior
unchanged, both verified against the code:
- The 'rough growth over-counts every content class' claim is false for
Cyrillic/Greek/Thai/Arabic (chars/4 vs ~2-3 chars/token on o200k):
growth there can under-count up to ~2x (#62605's direction). Document
the real backstops instead: an at/over-threshold real reading clears
the baseline (post-response gate fires on real usage within one call)
and the provider overflow handler compacts reactively.
- Document the two measurement bases (turn-prologue raw messages vs the
loop's fully assembled request that seeds the baseline) and why the
prologue's smaller basis can only OVER-defer — the loop's pre-API
pressure check re-runs the projection with the aligned basis before
every provider call, so a prologue over-defer never skips a needed
compaction.
The rough preflight estimate intentionally overestimates, but not by a
fixed margin: CJK text is counted at ~1.7x its o200k cost and
Responses-mode reasoning replay blobs at several times their billed
cost. Heavy sessions show rough estimates 2-3x real usage and compact
at 35-55% of the real window, stalling turns for minutes and discarding
detail (churn), because the defer guard only tolerated 5% rough growth
and sessions that never compressed had no baseline at all.
Pair every request's rough estimate (note_request_rough_estimate,
recorded in the conversation loop right after the pressure estimate)
with the provider's real prompt_tokens in update_from_response(), then
defer preflight while projected real usage — last real + rough growth
since that reading — stays under the threshold. Rough growth is itself
an overestimate of real growth, so the projection is an upper bound and
deferring below the threshold is safe; the provider's context-overflow
handler remains the backstop.
The baseline no longer ratchets on defer: it is refreshed by the
response pairing, and advancing it without a matching real reading
would shrink apparent growth and defer on stale data.
Sibling of the chat_completions zero-byte-args fix (previous commits):
a clean SSE close after content_block_start(tool_use) but before any
input_json_delta / message_delta yields an SDK final-message snapshot
whose content is NON-empty (the tool_use block is present, input={})
and whose stop_reason is None. That shape sailed past both
empty-stream guards (they only fire on empty content) and executed the
tool with empty input — no retry, no error: the same silent-data-loss
class as #80498, one provider transport over.
A legitimate completion always carries a stop_reason, so a
tool_use-bearing message without one is a mid-tool-call stream drop.
Raise EmptyStreamError for it, riding the same bounded stream-retry
(HERMES_STREAM_RETRIES) the eventless-stream case already uses.
Gate checked on both return paths (raw SDK snapshot and
accumulator-modified message). Regression tests cover the dropped
shape (mutation-verified: disabling the gate fails exactly that test),
the legitimate tool_use completion, and the text-only no-stop_reason
shape (pre-existing behavior preserved).
When the stream closes right after a tool call's name arrives but
before any argument bytes are delivered, has_truncated_tool_args
was never set (the existing check required a non-empty, whitespace-
stripped arguments buffer). The call fell through to a normal "stop"
finish_reason, later coerced to "{}" at dispatch and executed
silently with no arguments and no retry.
Route this case through the same dropped-mid-tool-call stub/retry
path already used for partially-truncated JSON.
Review follow-ups on the guard: state the accepted \n-residual in the
comment, reuse the span local in the next condition instead of
re-slicing, and short-circuit the substring checks before the regex.
Post-merge review of aecb9ca89 found the join guard over-broad: skipping
the join whenever ANY fragment self-matches _PREFIX_RE reopened a leak
for non-newline splits — sk-<15 chars>ESC<25 chars> masked only the
self-matching head and left the 25-char tail in cleartext (fully masked
before the guard; main never masked this shape at all, so the merged
state was still >= main, but the salvage's own coverage regressed).
Skip the join only when the span crosses a line boundary (\n / \r) —
that is the shape where adjacent legitimate text gets swallowed
(ghp_<token>-then-'button [ref=e3]' annotation bug). ESC/zero-width
controls never legitimately separate a token from prose, so joining
there is safe and restores full-tail masking.
Both legs mutation-checked: reverting to the unconditional skip fails
the new tail-mask test; removing the guard fails the annotation test.
CI slice 1/12 caught a regression in _mask_control_split_tokens: a
COMPLETE prefix token at end-of-line followed by ordinary text (browser
accessibility annotations: 'ghp_<tok>\nbutton [ref=e3]: Copy') was
joined across the newline into one stripped-copy match, and the mask
swallowed the adjacent line ('button' disappeared).
Join only when no fragment inside the span matches _PREFIX_RE on its
own — a self-matching fragment is already handled by the ordinary
prefix pass, so joining can only cause damage. All smuggling shapes
(ESC/ZWSP/newline splits with under-length fragments) still mask;
regression test added and mutation-checked (fails without the guard).
A masked secret's visible head/tail could carry control bytes (newline,
NUL, DEL, C1 0x80-0x9F, zero-width) into config/status/dump output.
Strip every control incl. \n/\t (display differs from redact_sensitive_text,
which preserves \n/\t as line structure) before slicing; all-control values
return the configured empty fallback.
Consolidates the previously-closed #58079 approach (strip controls before
masking) - supersedes it.
- Import file_safety._BLOCKED_PROJECT_ENV_BASENAMES instead of copying
it (comment-enforced parallel lists drift); lookup is now
case-insensitive to match file_safety's .lower() semantics (cat .ENV
on macOS/Windows case-insensitive filesystems reads the same secrets).
- Strip shell quotes plain split() leaves attached (cat ".env").
- Drop the dead _ENV_FILE_EXCLUDE_SUFFIXES logic (exact-basename
membership already excludes templates) and the stray blank-line noise.
- Document the defense-in-depth limits (sudo/full-path/substitution
readers) mirroring is_env_dump_command's precedent, and correct the
docstring overclaim about name-independence.
- Annotate command as str | None (tests pass None).
Terminal output from file-read commands (cat, head, tail, ...) uses
code_file=True, which skips the generic ENV-assignment redaction pass.
Reading a .env file through the terminal therefore leaked any key whose
value has no recognized vendor prefix (Mistral, Gemini AQ.*, tvly-dev-,
bu_, Spotify client secrets).
Detect file-read commands targeting .env-style basenames (mirroring
agent/file_safety's blocked list) and route them to code_file=False so
the existing ENV pass masks opaque values. Templates (.env.example,
.env.sample, ...) are excluded.
Salvaged from #61352 (145 commits of drift; conflict with the test-prune
wave resolved by NOT resurrecting pruned tests). Authored by @ShaoRou459.
Closes#61352
Review follow-up (W1): the pre-send transcript sanitizer
(agent_runtime_helpers.sanitize_tool_call_arguments) runs on the
PERSISTED messages list before every api_messages build and rewrites any
json.loads-failing argument string to "{}" in the transcript, prepending
a corruption marker to the paired tool result. That in-transcript repair
is deliberate (the stored turn must be replayable next call), but it
destroys the model's original bytes — for a truncated write_file call
those bytes are the user's streamed file content (#80498), and they
previously survived only as an 80-char log preview.
Until a sidecar-preservation design exists, make the bytes recoverable:
both destruction sites (the transcript sanitizer's WARNING and
_repair_tool_call_arguments' unrepairable-path WARNING) now log the full
original argument string bounded at 100KB instead of 80 chars. Corrupted
calls are rare; an oversized WARNING is a fair price for the only copy
of real user content.
The api_messages build used a shallow msg.copy(), decoupling only
top-level fields. Every nested container (tool_calls entries and their
function dicts, multimodal content-part lists, reasoning_details) stayed
aliased to the persisted history, so ANY in-place transform on the send
copy silently rewrote the stored transcript.
Probed every send-path transform against that aliasing shape on main:
content strip loop safe (top-level reassign)
_canonicalize_api_tool_calls (repair) LEAKED <- #80616's fix
_sanitize_messages_surrogates LEAKED (multimodal parts,
tc ids/args, reasoning)
_sanitize_messages_non_ascii LEAKED (multimodal parts)
_sanitize_api_messages safe
_drop_thinking_only_and_merge_users safe
The retry loop already believed the copies were independent - it
sanitizes messages AND api_messages separately (~L3555) - so the
aliasing was accidental everywhere.
Fix at the chokepoint: _clone_message_for_send clones every container
(dict/list) recursively while sharing immutable leaves, so every
downstream in-place transform - current and future - is safe by
construction. Cost is container-count, not string-bytes: 100KB argument
strings and base64 payloads are shared (measured ~0.5ms vs ~0.1ms per
1500-message build; noise next to one json round-trip). Same clone
applied to the prefill-message insert (same class, same pipeline).
The class-wide invariant test runs the full send-path transform
pipeline over an adversarial fixture (malformed args, surrogates,
non-ASCII, multimodal parts, reasoning fields) and asserts the history
stays byte-identical; an AST contract pins the build-site wiring so the
shallow copy can't quietly return. Both mutation-verified: reverting
the clone to shallow fails 4 isolation tests, unwiring the build site
fails the AST contract.
0xGr1mm's branch fix (previous commit) remains as defense in depth at
the exact site the #80498 incident hit; his regression tests and the
class-wide invariant give layered coverage.
`_canonicalize_api_tool_calls` promises copy-on-write in its own docstring
— "the persisted history is untouched" — and the call site repeats it:
"Operates on api_messages (the API copy) so the original conversation
history in `messages` is untouched."
The canonicalize branch keeps that promise (`tc = {**tc, "function": {...}}`).
The repair branch does not:
except Exception:
tc["function"]["arguments"] = _repair_tool_call_arguments(...)
`api_messages` is built with `msg.copy()` — a SHALLOW per-message copy — so
every `tool_calls` entry is the same dict object the persisted history
holds. Assigning into `tc["function"]` therefore writes through to the
stored turn. The sibling loop two lines above only touches `am["content"]`,
one level deep, which is why the aliasing never showed up there.
On the unrepairable path `_repair_tool_call_arguments` returns "{}", so
that write replaces the model's real arguments with an empty object in the
transcript. A stream that dies mid `write_file` loses the file content it
had already streamed — the reported symptom in #80498, where a chapter
draft was silently reduced to `{}` and only a WARNING remained:
Unrepairable tool_call arguments for write_file — replaced with empty
object (was: {"content": "# 骨架-第25章\n> 承接...)
Mirror the canonicalize branch: build a new tool-call dict instead of
assigning into the shared one. The API copy still carries "{}" — the
repair's whole purpose is to never ship broken JSON — but the history keeps
what the model actually sent, so the transcript, session persistence and
any later retry still have it.
The in-place write was not an oversight in isolation: it predates the memo
refactor, which preserved it deliberately for byte-parity. The existing
`test_history_not_mutated` asserts exactly this invariant but restricts
itself to valid arguments, and its docstring records the gap — "(Malformed
args take the in-place repair path — pre-existing behavior)". That is why
a test file whose header already claims "the persisted history is never
mutated (copy-on-write preserved)" stayed green through the bug.
Four tests close it: history keeps the original bytes, the send copy is
still repaired, a broken call does not disturb its siblings, and repeated
sends stay lossless. On unpatched main three of them fail; the parity and
complexity tests are unaffected because the difference is only observable
when the history list is separate from the send copy — which is the shape
production uses.
Refs #80498
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Thread request_provider through the sibling callsites that the
original PR left on resolved_provider: _set_relay_auxiliary_route
(observability metadata), the 'using X' log line, _is_anthropic_compat_endpoint
(Anthropic image conversion), _provider_requires_stream (streaming
detection), and the initial _relay_sync/async_completion calls.
These are NOT regressions (they passed 'auto' before), but widening
them ensures auto-routed calls to MiniMax get correct image conversion,
streaming-only providers are detected, and observability metadata
records the concrete backend instead of 'auto'.
Follow-up hardening on top of the salvaged #80809 watchdog, porting the
locked state-machine design from #75301 (credit: @Zeraphim):
- All lifecycle transitions (stale/cancelled/done) now happen under
request_client_lock. A user or monitor interrupt marks the request
'cancelled' so a racing stale timer can no longer misclassify the kill
as provider staleness and feed a false +1 into the #58962 cross-turn
circuit breaker.
- 'done' is set under the lock on completion, so a late timer callback
that lost the race to a successful response is inert instead of
leaving a spurious streak=1 behind the reset.
- Registration race closed: if the budget expires while the client is
still being constructed, _make_client aborts the freshly-registered
client and fails the call with a retryable TimeoutError instead of
opening a brand-new socket after the only watchdog already fired.
- _resolve_direct_stale_timeout now fails closed: a raising resolver
propagates (same as the worker path) instead of being swallowed into
an infinite budget that would silently disarm the watchdog and
reinstate the very hang #80759 is about.
4 new regression tests, each verified to fail against the pre-fix
watchdog implementation.
Co-authored-by: Zeraphim <diamantejc87@gmail.com>
Cron turns and delegated children are routed onto direct_api_call, which
ran the request inline with no stale detector. The abort plumbing was
registered but nothing ever invoked it, so a provider that accepted the
request and then went silent — connection held open, zero bytes, no
error — hung the run until an external actor killed it, which also
orphaned the execution row. The httpx read timeout is not a usable bound
(1800s default, and this failure mode never trips it), and the job-level
inactivity monitor was observed not to fire.
Arm a watchdog timer on the same budget the interrupt worker's poll loop
uses, so these turns get exactly the patience every other non-streaming
request already gets. On expiry it only aborts the in-flight sockets
through the already-registered hook — it never issues a request, so the
inline / no-worker property that fixes the nested-pool deadlock is
preserved — bumps the cross-turn stale circuit breaker, and surfaces a
retryable TimeoutError so the outer loop reconnects on a fresh pool.
Fixes#80759
Review follow-up: after the walk-back widening, the exempted assistant
is often not the final message, and the partial-batch shape is
byte-identical to a settled-but-malformed orphan — so say 'presumed
pending' and document WHY presuming is safe (sanitize_api_messages
step 2 stubs any genuinely unanswered call pre-API on every path).
Widen #79293's trailing-in-flight guard from 'last message is assistant'
to 'last non-tool message is assistant': a multi-call batch snapshotted
between the executor's per-result appends looks like
[..., assistant(c1,c2,c3), tool(c1)] — c2/c3 are pending, not orphaned,
but the tail-only guard missed that shape and stripped them (same silent
result loss as the original bug, via concurrent /compress or the gateway
hygiene pass).
Preserving is safe on both shapes: the pre-API chokepoint
(sanitize_api_messages step 2) injects stub results for any call that
genuinely never gets an answer, while stripping a live call silently
loses its late result.
test_sanitizer_strips_orphaned_keeps_valid's mixed valid/orphan shape
moves mid-list — at the tail it is byte-identical to a live partial
batch and the sanitizer now correctly presumes in-flight there.
New regression test fails without the walk-back (c2/c3 stripped),
passes with it.
Tool_executor.py appends role=tool results AFTER running each call. When
context compression fires mid-chain, the trailing assistant(tool_calls)
message is a pending request whose result has not yet been appended.
_sanitize_tool_pairs previously stripped it as an 'orphan', so when the
executor later appended the real result, repair_message_sequence dropped
it as unmatched and the completed side effect (and final synthesis) was
lost. Preserve the trailing in-flight call verbatim; only genuinely
orphaned calls in the discarded region are stripped.
Adds regression tests: three unit tests for _sanitize_tool_pairs plus an
end-to-end test reproducing compression -> side-effect completion ->
result-returned flow. Confirmed failing on pre-fix code, passing with
the fix.
Adversarial review of the salvaged recovery found a reachable fail-open:
compression continuations inherit the rotated agent's model_config
verbatim (publish_compression_child callers pass
agent._session_init_model_config), so a delegate subagent's continuation
carries _delegate_from=<the delegate's own parent>. The marker-PRESENCE
filters in reopen_orphaned_compression_session and
find_live_compression_child misclassified such a REAL continuation as a
delegate child:
- reopen: parent 'orphaned' -> reopened while a live continuation exists
-> two live heads in one lineage (verified with a live repro)
- find_live: adoption misses the continuation (fail-closed, masked the
fork pre-PR; the PR made it active)
Fix: markers only disqualify a child when they point at the queried
parent (shared _NON_CONTINUATION_CHILD_FILTER_SQL fragment, also
resolving the duplicated-SQL drift risk flagged by the reuse reviewer).
Both directions regression-tested: reopen fails closed on an
inherited-marker continuation; find_live adopts it.
Also from review: reopen-failure log raised debug->warning (the failure
hard-fails the turn moments later), commit-semantics hardening comment
on the lease DELETE path, blank-line nit.
The three read-only projection walks (get_compression_tip,
list_sessions_rich chain, resume walk) share the marker-presence shape
but fail closed (skip a continuation -> resume shows the parent), and
the fixed adoption path self-heals that case at turn start; left as-is.
Inspired by virgiliojr94/book-to-skill (MIT): /learn now picks the skill
shape by the source. Workflows and small sources still get one tight
SKILL.md; books, paper stacks, specs, and large doc corpora get a
knowledge-base layout — a lean always-loaded SKILL.md index plus one
distilled file per chapter/topic under references/, loaded on demand via
skill_view so query cost stays proportional to the answer.
- agent/learn_prompt.py: new _KNOWLEDGE_SKILL_STANDARDS block (index +
per-chapter references/, structure-not-summary distillation, never
reproduce source passages, fold-in instead of duplicating) and a
_SOURCE_HYGIENE block pinning extracted source text as data and
dropping invisible/bidi Unicode (Trojan Source class). Clarified that
the ~200-line cap and hub-skill ban apply to SKILL.md itself, not a
knowledge skill's own references/ files.
- tests: contracts for the knowledge-base layout, the three embedded
standards blocks, and the source-hygiene coverage.
- docs: skills.md documents the knowledge-base shape.
The native Responses stream does carry summary_index, so the part boundary is
structured data here rather than something to infer. Break on a change of
index, and leave streams that send no index (plain reasoning_text) untouched.
Reasoning-summary models emit one reasoning_content delta per completed
summary part, each a self-contained bold heading. The Responses API delimits
those parts with summary_index; the OpenAI chat wire carries no such field —
verified live against Nous Portal, whose reasoning chunks contain nothing but
delta.reasoning_content — so concatenating them glued every part into one
unspaced, half-bold paragraph.
Re-derive the boundary from the signal the wire does carry: a delta opening a
closed bold heading against a mid-line tail. This matches Hermes own Responses
adapter, which already joins its summary parts with a blank line.
- Single source for the approval-derived bound: public human_wait_ceiling()
in tools/approval.py; the gate's lock-timeout helper delegates to it
instead of re-deriving timeout + margin (was duplicated in two modules
and reached for a private _get_approval_timeout).
- Shared _clamped_window_seconds() for the close-time accrual and the
open-window read, so the two clamps are identical by construction.
- Gate __init__ grows session_key kwarg; tests construct via the real
constructor instead of mutating privates post-hoc.
- Gateway test resolves its pending approval via resolve_gateway_approval()
(the production /deny path) instead of hand-rolling queue-entry internals.
- Docstring accuracy: human_wait_seconds monotonicity caveat under cap
eviction; s/pre_tool_block/pre_tool_call/ hook name.
Review-driven follow-up to the #79719 fix:
- Clamp the CLOSE-side accrual too: a wedged window that eventually closed
used to inject its full unclamped overstay into completed_seconds,
retroactively extending a running batch's deadline by hours. Both clamps
now share one ceiling helper (_human_wait_ceiling = approvals.timeout +
HUMAN_WAIT_MARGIN_S), and the gate's lock-timeout uses the same margin
constant so the bounds cannot drift apart.
- Evict idle sessions until the table is under the cap (was: at most one
per insert, so churn could outgrow _HUMAN_WAIT_MAX_SESSIONS). Entries
with an open window are still never evicted.
- Log (debug) instead of silently swallowing a failed session-key snapshot
in the gate constructor.
Tests: close-side clamp regression + table-cap assertion added; suite at
17 passed.
A tool wedged inside _ConcurrentToolAuthorizationGate hung the whole turn
forever (#79719): excluded_seconds() measured residency in gate.run() —
arbitrary code — so an open window grew 1:1 with wall clock and the batch
deadline's remaining was constant (remaining = deadline - window_started;
now cancels out). A hanging pre_tool_call plugin or an approval round-trip
to a dead client defeated the deadline entirely. The serialization lock was
also an unbounded acquire, so every other worker needing authorization
parked behind the wedged holder forever.
Fix, in two halves:
- tools/approval.py grows per-session human-wait accounting
(human_wait_window / human_wait_seconds). The two places that are
verifiably blocked on a HUMAN — the CLI approval prompt and the gateway
approval poll loop — mark their own windows. Both are intrinsically
bounded by approvals.timeout; the open-window read is additionally
clamped to that timeout plus a margin as belt-and-braces.
- _ConcurrentToolAuthorizationGate keeps only serialization, with a bounded
acquire (approvals.timeout + 60s; on expiry the prompt runs unserialized —
the same degradation the start-order gate accepted in #79705).
excluded_seconds() becomes a baseline-delta read of the session's
human-wait total.
A wedged plugin now contributes nothing to the exclusion, so the batch
times out at the normal deadline with correctly labeled results, while a
genuine approval wait — which can legitimately exceed any fixed bound —
still extends the deadline in full. E2E (real AIAgent, worktree imports):
wedged-plugin batch on main never ends (>30s observed, 3s deadline); with
the fix it ends at 3.0s. A 4s simulated approval over a 2s deadline
completes without a timeout label.
Closes#79719
/refine [focus] fires the existing background review fork
(AIAgent._spawn_background_review) immediately instead of waiting for
the automatic 10-turn memory / 10-iteration skill nudge counters.
Optional focus instructions are appended to the review prompt so the
fork prioritizes what the user asked for (e.g. '/refine save the
deploy workflow as a skill').
- New optional focus parameter threaded through
_spawn_background_review -> spawn_background_review_thread.
Automatic post-turn reviews pass None and their prompts are
byte-identical to before.
- CLI handler snapshots conversation_history; gateway handler pulls
the idle session's cached AIAgent from _agent_cache (rejected while
the agent is running).
- Review runs in a daemon thread against the snapshot — live
conversation, message alternation, and prompt cache untouched.
- Slack stays under the 50-slash cap via /hermes refine.
Adapted from the /refine concept in Prime Intellect's Prime-Agent
(Continual Harness); Hermes' equivalent durable state is the
memory + skill stores, so the review fork is the natural target.
/heartbeat every <interval> <prompt> gives the current session one
recurring instruction. When the session is idle and the interval has
elapsed, the prompt is injected as a plain user turn — same
conversation, same context, prompt cache and role alternation
untouched.
- CLI: idle-poll watchdog thread (wake-word watchdog pattern) feeding
_pending_input; gateway: single gateway-wide async poller injecting
through the adapter FIFO. Busy sessions coalesce their tick to the
next idle poll.
- Missed ticks coalesce (anchor resets on fire) — a busy hour yields
ONE heartbeat turn, never a backlog. Real user messages always win.
- 60s interval floor; injected prompt carries a don't-invent-work
guard so idle heartbeats don't generate busywork.
- State persists in SessionDB.state_meta (heartbeat:<session_id>),
survives /resume, migrates across compression session rotations
alongside /goal state.
- Session-scoped and in-process by design — durable cross-process
schedules remain the cron subsystem's job (docs draw the boundary).
- Slack stays under the 50-slash cap via /hermes heartbeat; ghost-text
suggester now prefers the shortest prefix match so /he still
suggests /help.
Adapted from the session-heartbeat concept in Prime Intellect's
Prime-Agent (/heartbeat).
Port from google-gemini/gemini-cli#28700: when an interrupted/failed turn
leaves history ending on an unanswered tool result and the user sends a new
message, fusing the two into one Gemini user content makes the model read the
trailing text as a continuation of the tool result — it 'finishes your
sentence' instead of answering.
Builds on #68863 (@rille111), which split the mixed functionResponse/text
merge but emitted two consecutive user contents — a shape Gemini's
alternation contract rejects with HTTP 400 on other request paths (#55125).
This follow-up interposes gemini-cli's INTERRUPTED_RESPONSE_PLACEHOLDER model
turn between the split contents so the request stays alternation-valid while
the user's message remains a turn of its own.
Do not fold a human user text turn into a preceding functionResponse
user content. Gemini 3 accepts that fold with HTTP 200 but then returns
an empty model response.
Contract:
- ordinary same-role merges remain (parallel tool results, back-to-back
plain user texts) for Gemini alternation
- only mixed functionResponse/text user turns are split
A gateway session whose summary model keeps timing out no longer retries
compaction on the same fixed interval forever.
The in-agent compressor already escalates repeat summary timeouts
60 -> 300 -> 900s (ContextCompressor.record_timeout_failure), but that ladder
reads the in-memory _consecutive_timeout_failures counter and
bind_session_state() zeroes it (context_compressor.py:1645). Session hygiene
constructs a FRESH AIAgent for every run (gateway/run.py:16820) and re-binds
state each time, so from the gateway that streak is structurally always 0 --
only the flat hygiene_failure_cooldown_seconds (300s) could ever be recorded.
Issue #79624 reported exactly that steady state: an oversized session
(1053 messages, ~119.5k tokens) whose aux model always timed out, re-attempting
compaction every 300s across five days until the reporter deleted the session
by hand.
Track the streak on PersistentState instead, which outlives the per-run agent
and is not cleared by turn/boundary resets, so consecutive hygiene failures
climb 300 -> 900 -> 2700s and then saturate. Both failure sites (progress
timeout and aborted compression) feed it; a real compression resets it, so a
session that recovers starts from the first rung again. The ladder multiplies
the configured base, so operators who tuned
hygiene_failure_cooldown_seconds keep their first rung. Per-session, so one
wedged chat cannot penalize other conversations.
Deliberately NOT changed, since each is a maintainer policy call rather than a
defect (all three are written up on #79624):
- no durable failure-streak column, so escalation still resets on restart
- the gateway 30s / in-agent 120s / aux-client 300s-floor timeout mismatch
- no `hermes doctor` check or `hermes sessions list` marker for a session
stuck in a compression-failure cooldown
Note the reported exit(1) is NOT a crash: it is the deliberate
_signal_initiated_shutdown path (gateway/run.py:26746-26751, #5646) that lets
systemd Restart=on-failure revive the gateway after a bare SIGTERM, and it
fires on every `systemctl restart` independently of compaction. The compaction
log lines appear after the shutdown line because the gateway-owned executor is
torn down with shutdown(wait=False, cancel_futures=True) (run.py:21164), so an
in-flight turn keeps logging during teardown. Full analysis on the issue.
Post-review hardening (Phase 2c + /simplify-code found five real defects in the
first cut):
- the recovery gate hand-rolled `_new_tokens < _approx_tokens` when a canonical
predicate already existed: `compression_made_progress` (agent/turn_context.py,
#39548). They disagree on 3 of 5 cases -- the hand-rolled form misses a
row-count win when the summary keeps the token estimate flat, misses one
where the summary is slightly MORE verbose (so a genuinely recovered session
would keep escalating forever), and counts a sub-5% wobble as recovery. Now
reuses the shared predicate, promoted from `_compression_made_progress` to a
public name with the old private name kept as a back-compat alias so the
existing importer (tests/agent/test_protected_tail_pressure_61932.py) and any
patcher of that symbol keep working.
- the reset was gated on "not aborted", but the degenerate "did not rotate or
compact in place" branch (#21301) is NOT aborted and yields zero reduction,
so a session wedged there reset its streak every run and could never
escalate -- silently defeating the fix. Now gated on real progress.
- no absolute ceiling: base * 9 reaches 9h at an operator base of 3600s,
indistinguishable from "compaction switched off". Added
_HYGIENE_COOLDOWN_MAX_SECONDS = 3600, mirroring the in-file
_RECONNECT_BACKOFF_CAP precedent.
- the reset used the get-or-create accessor to write a 0 that was already 0,
materialising a _sessions entry (never evicted). Now peeks.
- the abort verdict was probed twice, leaving the reset/record mutual
exclusion implicit; a future await between the probes would have broken it
silently. Computed once into _hyg_aborted.
Tests: 19 new in tests/gateway/test_hygiene_failure_cooldown_ladder.py --
ladder escalation, saturation, the absolute cap, per-session isolation,
reset-on-recovery, custom/zero base, PersistentState scoping (a mutation moving
the field to TurnState fails), degraded runners, the progress gate, the exact
progress-predicate semantics the gate depends on, and end-to-end that the
escalated value is what reaches the state DB. All 12 mutations caught, including
ones that restore the flat cooldown (the original bug), ungate the reset, swap
the canonical predicate back for the hand-rolled comparison, remove the cap, and
share the streak globally; the harness hard-errors when a mutation cannot be
applied, since a silently no-op mutation check is worse than none -- an earlier
version of it WAS silently no-opping after a refactor. The gate's contract test
slices by AST node span rather than a fixed character count, which had already
truncated once as the block grew. gateway hygiene + session-state + the three
touched agent compression suites: 50 passed; ruff clean.
E2E with real imports demonstrates the premise rather than asserting it:
bind_session_state zeroes the in-agent counter, and the recorded deadlines go
300 -> 900 -> 2700 -> 2700 -> 2700s where they were previously a flat 300s.
Reported by @yucezerey (#79624), whose state.db column dump and
"deleting the session fixed it" datapoint made the real mechanism findable.
Follow-up to the salvaged start-order gate bound. Two gaps remained, both
reachable through the same knob.
1. The gate bound ignored the batch deadline it sits under. With
HERMES_CONCURRENT_TOOL_TIMEOUT_S below 120s the deadline fired first, so
the parked tools were still reported as "timed out" without ever running --
the exact bug the bound exists to fix. The gate now clamps to
min(120s, batch_timeout / 2), matching the sibling constant's documented
habit of relating the two timeouts.
2. A gate-parked worker released purely by its own timeout could wake up after
the batch was abandoned and dispatch its tool anyway: wasted work whose
result nobody reads, a duplicate post_tool_call for a tool_call_id the turn
already closed as timeout, and agent._current_tool left pointing at a dead
tool for the rest of the session (the main thread's reset already ran).
Abandonment is now a first-class wakeup: both abandon sites set an event and
notify the condition, and a released worker raises _BatchAbandoned instead
of dispatching. Parked threads are reclaimed in milliseconds rather than one
full gate timeout plus a tool runtime.
Also names the tool in the gate-timeout warning. The closure's function_name
binds the last-parsed tool, so logging it directly would have printed the wrong
name; it is threaded through _begin_in_order instead.
Measured, 3-tool batch with the first tool wedged during dispatch:
main PR as-is with this commit
dispatched in batch 0 0 tool_b, tool_c
dispatched after return 0 2 (ghost) 0
_current_tool leaked no "tool_b" no
Adds tests/run_agent/test_start_order_gate.py (3 tests). Mutation-checked
against the parent commit: the starvation guard passes there (it binds the
salvaged fix), while the deadline-clamp and abandonment guards both fail,
reproducing the ghost dispatch as
"tool(s) dispatched after the batch was abandoned: [tool_a, tool_b]".
_begin_in_order parks each concurrent tool worker on a timeout-less
Condition.wait_for until every earlier-ordered tool has advanced through
its dispatch. If one tool wedges during dispatch (observed in production:
a stuck skill_view; also reproducible via any blocking authorization),
three failures compound: every later-ordered worker is starved and never
starts; the batch deadline then falsely reports those never-started tools
as "timed out" (sub-second read_file/search_files calls get blamed while
having done zero work, and the model reasons against that false failure
info); and after the batch is abandoned the parked workers leak forever —
f.cancel() cannot cancel running threads, the per-thread interrupt flag
is never polled inside wait_for, and nothing notifies the condition
again. Confirmed with a faulthandler all-threads dump taken after batch
abandonment showing workers still parked at the gate.
Bound the wait at 120s; on expiry, log a warning and proceed out of
order (worst case: interleaved approval prompts — strictly better than
permanent starvation). The >= predicate lets one worker's timeout-jump
release every skipped worker immediately, and max() keeps the counter
monotonic for out-of-order advancement.
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
A kanban worker that fires a cron job in-process no longer leaks its task
identity into the cron agent.
The worker is a normal `hermes chat -q` CLI agent whose default toolset
includes `cronjob`, running with HERMES_KANBAN_TASK legitimately set in its
own environment. `cronjob(action="run")` calls run_one_job() -> run_job()
in that same process, so the cron AIAgent was misidentified as that worker:
kanban toolset force-added, kanban-worker protocol injected into its system
prompt, and kanban_complete defaulting task_id to $HERMES_KANBAN_TASK --
letting an unrelated cron job close the worker's task and overwrite real
results.
Fixed with a ContextVar (`non_dispatcher_owned_context`), not by clearing
os.environ. The env is process-global and shared with three concurrent
readers that all need the real values:
* the worker's own claim heartbeat -- run_agent._touch_activity ->
heartbeat_current_worker_from_env reads TASK/CLAIM_LOCK/RUN_ID, and the
cron-run heartbeat thread drives it every 10s. Clearing them silently
no-ops the heartbeat, so after DEFAULT_CLAIM_TTL_SECONDS (15 min) the
dispatcher reclaims a task whose worker is still alive and re-dispatches
it -- the same duplicate-work failure from the other direction.
* the gateway's kanban watchers, which do their own HERMES_KANBAN_BOARD
save/restore around a slow decompose_task() LLM call.
* concurrent cron jobs, which take a *shared* read lock
(_terminal_cwd_lock.acquire_read) and so interleave: job A clears, job B
snapshots empty, A restores, B clears and its restore no-ops -- the
worker's identity is destroyed permanently.
`is_dispatcher_owned_worker_context()` is now the single predicate every
HERMES_KANBAN_* identity gate consults before trusting those vars. It also
closes a pre-existing gap in agent/skill_utils.py, which read the vars
without consulting the delegate_task ContextVar at all; the `kanban` verdict
additionally bypasses _ENV_DETECT_CACHE, since a context-dependent answer
must not be memoized process-wide.
HERMES_KANBAN_BOARD/DB/WORKSPACES_ROOT are left untouched, so the #20074
board pin and the dispatcher's path overrides keep working.
Tests: 18 new, including thread-isolation, concurrent-cron-jobs, and an AST
invariant over _default_spawn that fails if the dispatcher gains a var that
is neither identity-gated nor explicitly classified behaviour-only. All six
mutations are caught, including one that reintroduces the os.environ clear.
tests/cron/ + kanban suites 440 passed; model_tools/skill_utils/boards 63
passed; ruff clean.
Reported and diagnosed by Geoff Friesen (#78961), who identified the symptom
and the exact gating mechanism.
Co-authored-by: Geoff Friesen <gfriesen1@users.noreply.github.com>
Teknium
0xarkstar
kshitij
Yishova
HexLab98
Alan Hsu
kshitij
joaomarcos
Soheil Fakour
Peter
Gr1mmJ4w
Gille
Shotflame
izumi0uu
brooklyn!
Rickard Robin
Sylvain Baeriswyl