## Thinking Path
> - Paperclip is the open source app people use to manage AI agents for
work
> - Hermes adapter produces terminal output with ANSI color codes on
stdout
> - These escape sequences flow through the UI parsers untouched and
render as raw garbage text
> - This PR adds ANSI stripping at the entry point of all four Hermes
parse-stdout entry points
> - The same regex is already proven in claude-local adapter
> - The benefit is clean, readable terminal output for Hermes agents
## Linked Issues or Issue Description
No existing issue. This is a bug report:
**What happened**
Hermes terminal output displayed ANSI color codes as raw text in the
Paperclip UI, making agent output unreadable.
**Expected behavior**
Terminal output in run transcripts should be clean text without
invisible control characters.
**Steps to reproduce**
1. Connect a Hermes agent to Paperclip
2. Create and assign a task to the agent
3. View the run transcript — ANSI escape codes appear as raw garbage
**Paperclip version or commit**
e6407b322 (upstream master)
**Deployment mode**
local_trusted
## What Changed
- Added `stripAnsi()` function using the same regex pattern from
claude-local adapter (quota.ts) — strips CSI and
OSC sequences
- Applied at entry point of `parseHermesStdoutLine` in hermes_local (TS
+ CJS)
- Applied at entry point of `parseHermesGatewayStdoutLine` in
hermes_gateway (TS + CJS)
- CJS files keep the function inline since the dynamic parser sandbox
has no module loader
- 5 files changed, +123/-8 lines
## Verification
- Smoke tested with real ANSI patterns from Hermes output — all samples
pass
- `pnpm --filter @paperclipai/hermes-paperclip-adapter exec vitest run
src/ui/parse-stdout.test.ts` — 9 passed
- TypeScript compiles clean for both hermes and hermes-gateway packages
- Adapter tests pass (5/6, 1 pre-existing Windows CI failure unrelated)
- Live tested on running Paperclip instance — ANSI codes no longer
appear in transcripts
## Risks
Low risk. Only affects Hermes parser output. Regex already proven in
claude-local adapter. No logic changes to parse
behavior — only strips invisible control characters before parsing.
## Model Used
DeepSeek V4 Pro — reasoning mode, tool use
## Checklist
- [x] I have included a thinking path that traces from project context
to this change
- [x] I have specified the model used (with version and capability
details)
- [x] I have checked ROADMAP.md and confirmed this PR does not duplicate
planned core work
- [x] I have searched GitHub for duplicate or related PRs and linked
them above
- [x] I have either (a) linked existing issues with `Fixes: #` / `Closes
#` / `Refs #` OR (b) described the issue
in-PR following the relevant issue template
- [x] I have not referenced internal/instance-local Paperclip issues or
links (only public GitHub `#NNN` /
`github.com/paperclipai/paperclip` URLs)
- [x] My branch name describes the change (e.g. `docs/...`, `fix/...`)
and contains no internal Paperclip ticket id
or instance-derived details
- [x] I have run tests locally and they pass
- [x] I have added or updated tests where applicable
- [ ] I have updated relevant documentation to reflect my changes
- [x] I have considered and documented any risks above
- [ ] All Paperclip CI gates are green
- [ ] Greptile is 5/5 with no open P2s, recommendations, or follow-ups
- [x] I will address all Greptile and reviewer comments before
requesting merge
70c86d2c73