[codex] Require complete main-agent skill reads#27044
Conversation
|
All contributors have signed the CLA ✍️ ✅ |
0274429 to
fad5259
Compare
|
I have read the CLA Document and I hereby sign the CLA |
There was a problem hiding this comment.
Pull request overview
This PR strengthens the “How to use skills” prompt text in core-skills to require the main agent to fully read selected SKILL.md files (and any task-required referenced instruction files) through EOF, and to prohibit delegating interpretation of skill instructions to subagents—while still allowing subagents to perform task work when permitted by the skill.
Changes:
- Update skill-usage instructions (both absolute-path and alias-root variants) to require complete
SKILL.mdreads and explicit continuation on truncation/pagination until EOF. - Require the main agent to personally read task-required referenced instruction/reference files and prohibit delegating skill-instruction interpretation to subagents.
- Add a unit test asserting these instruction invariants and guarding against reintroducing the older “read only enough” wording.
💡 Add Copilot custom instructions for smarter, more guided reviews. Learn how to get started.
There was a problem hiding this comment.
💡 Codex Review
Here are some automated review suggestions for this pull request.
Reviewed commit: fad5259746
ℹ️ About Codex in GitHub
Your team has set up Codex to review pull requests in this repo. Reviews are triggered when you
- Open a pull request for review
- Mark a draft as ready
- Comment "@codex review".
If Codex has suggestions, it will comment; otherwise it will react with 👍.
Codex can also answer questions or update the PR. Try commenting "@codex address that feedback".
| - Missing/blocked: If a named skill isn't in the list or the path can't be read, say so briefly and continue with the best fallback. | ||
| - How to use a skill (progressive disclosure): | ||
| 1) After deciding to use a skill, open its `SKILL.md`. Read only enough to follow the workflow. | ||
| 1) After deciding to use a skill, the main agent must open and read its `SKILL.md` completely before taking task actions. If a read is truncated or paginated, continue until EOF. |
There was a problem hiding this comment.
Bound complete skill-file reads
Long or paginated user/plugin SKILL.md files are now read until EOF, letting arbitrary skill content enter model context without a hard cap. That conflicts with Model visible context; keep a bounded read/truncation path.
Useful? React with 👍 / 👎.
Summary
SKILL.mdfiles completely, continuing truncated or paginated reads through EOFWhy
Partial reads can skip routing and verification requirements later in skill instructions. Delegated summaries can also omit constraints the main agent needs to follow. The existing "Read only enough" wording made both behaviors appear acceptable.
Impact
Agents should follow complete selected skill instructions while continuing to avoid unrelated references, scripts, and assets. Subagents remain available for task execution where permitted.
Test plan
just test -p codex-core-skills(101 passed)just fmtgit diff --check