docs: clarify large artifact context tooling#114
Conversation
📝 WalkthroughWalkthroughThis pull request adds a single instruction to AGENTS.md that permits optional use of compression-based navigation tools for extremely large non-source artifacts such as CI logs or debug traces, contingent on verifying conclusions against the original uncompressed artifact before making code changes or final assertions. ChangesAgent Guidance Documentation
🎯 1 (Trivial) | ⏱️ ~2 minutes 🚥 Pre-merge checks | ✅ 5✅ Passed checks (5 passed)
✏️ Tip: You can configure your own custom pre-merge checks in the settings. ✨ Finishing Touches🧪 Generate unit tests (beta)
Thanks for using CodeRabbit! It's free for OSS, and your support helps us grow. If you like it, consider giving us a shout-out. Comment |
![coderabbitai[bot]](https://avatars.githubusercontent.com/in/347564?s=60&v=4){kind=small}
There was a problem hiding this comment.
Actionable comments posted: 1
🤖 Prompt for all review comments with AI agents
Verify each finding against current code. Fix only still-valid issues, skip the
rest with a brief reason, keep changes minimal, and validate.
Inline comments:
In `@AGENTS.md`:
- Line 103: Update the AGENTS.md line that references "chopratejas/headroom" to
describe it as a compression/token-reduction tool (e.g., library/proxy/MCP
server for compressing tool outputs/logs) used to reduce tokens before
processing, rather than calling it a "navigation aid"; keep the existing
guardrail language requiring agents to verify conclusions against the original
uncompressed artifact before editing code or making final claims, and ensure the
sentence mentions examples like CI logs, debug traces, generated JSON, or
massive tool output for clarity.
🪄 Autofix (Beta)
Fix all unresolved CodeRabbit comments on this PR:
- Push a commit to this branch (recommended)
- Create a new PR with the fixes
ℹ️ Review info
⚙️ Run configuration
Configuration used: Path: .coderabbit.yaml
Review profile: CHILL
Plan: Pro Plus
Run ID: 53ff238c-c9bb-43a5-b0df-c472cd40c80f
📒 Files selected for processing (1)
AGENTS.md
| @@ -100,6 +100,7 @@ Before working on certain areas, read the relevant protocol doc to avoid mistake | |||
| - When debugging CI failures, download the GitHub Actions artifacts from the failed run. The most useful are: (1) `per-test-logs-*` — contains per-test-file `<stem>.stdout.log` and `<stem>.stderr.log` mirroring the test path (e.g. `node/pkc/pkc.stderr.log`); start with the stderr log of the failing test file. (2) `test-server-log-*` — the full `test_server.log` with community-side DEBUG output. (3) `test-logs-*` — aggregate stdout/stderr from `--log-prefix`. (4) `vitest-*` — vitest JSON reports. | |||
| - To capture stdout/stderr from `run-test-config.js` to log files, use `--stdout-log <path>` and `--stderr-log <path>`, or use `--log-prefix <prefix>` to automatically create `<prefix>.stdout.log` and `<prefix>.stderr.log`. DEBUG output (from the `debug` module) goes to stderr. | |||
| - To capture per-test-file logs, use `--per-test-logs <dir>`. Each test file gets its own `<stem>.stdout.log` and `<stem>.stderr.log` under a directory structure mirroring the test path (e.g. `test/node/pkc/pkc.test.ts` → `<dir>/node/pkc/pkc.stderr.log`). This also captures DEBUG output by redirecting the `debug` module through `console.error`, which vitest then captures per-test via `onUserConsoleLog`. Example: `DEBUG="pkc-js*" node test/run-test-config.js --pkc-config "local-kubo-rpc,remote-pkc-rpc" --per-test-logs .tmp/test-logs test/node/pkc/pkc.test.ts`. | |||
| - For very large non-source artifacts such as CI logs, debug traces, generated JSON, or massive tool output, agents may optionally use `chopratejas/headroom` or a similar local compression tool as a navigation aid, but must verify conclusions against the original uncompressed artifact before editing code or making final claims. | |||
There was a problem hiding this comment.
🧩 Analysis chain
🏁 Script executed:
#!/bin/bash
# Description: Verify if the referenced tools exist as GitHub repositories
echo "Checking chopratejas/headroom..."
gh repo view chopratejas/headroom --json name,description,url 2>&1 || echo "Repository not found"Repository: pkcprotocol/pkc-js
Length of output: 320
Clarify chopratejas/headroom wording to match its purpose (AGENTS.md line 103).
chopratejas/headroom exists and is a library/proxy/MCP server for compressing tool outputs/logs (token reduction). Adjust the guidance from “navigation aid” to “compression/token-reduction (e.g., via headroom) before processing,” while keeping the “verify against the original uncompressed artifact” guardrail.
🧰 Tools
🪛 LanguageTool
[style] ~103-~103: As an alternative to the over-used intensifier ‘very’, consider replacing this phrase.
Context: ...-logs test/node/pkc/pkc.test.ts`. - For very large non-source artifacts such as CI logs, d...
(EN_WEAK_ADJECTIVE)
🤖 Prompt for AI Agents
Verify each finding against current code. Fix only still-valid issues, skip the
rest with a brief reason, keep changes minimal, and validate.
In `@AGENTS.md` at line 103, Update the AGENTS.md line that references
"chopratejas/headroom" to describe it as a compression/token-reduction tool
(e.g., library/proxy/MCP server for compressing tool outputs/logs) used to
reduce tokens before processing, rather than calling it a "navigation aid"; keep
the existing guardrail language requiring agents to verify conclusions against
the original uncompressed artifact before editing code or making final claims,
and ensure the sentence mentions examples like CI logs, debug traces, generated
JSON, or massive tool output for clarity.
1 participant
Summary
Verification
No build or test run: documentation-only AGENTS.md change.
Summary by CodeRabbit