Repository navigation
docs: add an agent playbook and keep agent docs in sync with action.yml - #27
Merged
Merged
Conversation
Co-Authored-By: Claude Sonnet 5.5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01HxpS9r6qDihN2RKxQCwU2R
- tests/check-action.py now parses every yaml block in the docs, fails when a call to this Action uses an input that is not in action.yml or is not pinned to @v1, and checks relative links and #anchors in the docs - the metadata workflow runs when AGENTS.md, llms.txt or docs/ change - the playbook tells agents to read the failure annotation first, which now carries the first cause line (#26) Co-Authored-By: Claude Sonnet 5.5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01HxpS9r6qDihN2RKxQCwU2R
Rendered emailsThis PR rendered email templates.
Open the run → Artifacts → download → open |
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
2 participants
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
A better explanation of this Action for AI agents
AGENTS.mdmixed two audiences (agents adding the Action to a repository, and agents changing this repository) and described the Action, but did not give an agent a procedure to follow. This PR adds one.What's new
docs/agent-playbook.md: a four-step procedure (inspect the repository, choose a render mode, write the workflow, verify), a decision flowchart, eleven recipes (script, other flag styles, MJML/pandoc, tools not on the runner, catalog command, pnpm/yarn/bun, monorepo, matrix, private packages,fail-fast: false, two comments, scheduled run), a table that maps each log message to a fix, and a list of hard rules. It tells agents to read the failure annotation first, which now carries the first cause line (feat: include the failing command's cause in the error annotation #26).AGENTS.mdopens with a "Start here" section: where to go for each audience, and the five rules that matter most. The long reference sections are unchanged.llms.txtand the README link to the playbook; CHANGELOG updated.So the docs cannot drift
tests/check-action.pynow also:```yamlblock in the README,AGENTS.md,llms.txt, CHANGELOG anddocs/;examples/uses an input that is not inaction.yml, or is not pinned to@v1;#anchorspoint at files and headings that exist.I checked the checks: injecting a bad input name, an
@mainpin, a missing file, a dead anchor and broken YAML into a scratch copy failed on each one. The metadata workflow now also runs whenAGENTS.md,llms.txtordocs/change.Testing
All local suites pass (31 + 102 + 13 assertions, plus the static checks). No behavior of the Action changes in this PR.
🤖 Generated with Claude Code
https://claude.ai/code/session_01HxpS9r6qDihN2RKxQCwU2R
Generated by Claude Code