Adopt Claude Code agent conventions#40
Conversation
Rename the Codex-authored agent config to its Claude Code equivalents and refresh the skills against the Opus 4.8 prompting guidance. - Rename AGENTS.md to CLAUDE.md and .agents/ to .claude/. - Delete the Codex-only openai.yaml manifest; SKILL.md frontmatter already carries name, description, and invocation. - Switch skill invocation to slash syntax (/new-feature, /feature-flags) and point the new-feature artifact instructions at CLAUDE.md. - Rewrite the create-release, feature-flags, and new-feature skills for literal-scope wording, positive framing, and the reasoning behind each non-obvious rule, preserving every command and fact. Product-level Codex references in README.md and docs/ are real MCP integration features and stay untouched. Co-authored-by: Claude <noreply@anthropic.com> Co-authored-by: GitButler <gitbutler@gitbutler.com>
![gemini-code-assist[bot]](https://avatars.githubusercontent.com/in/956858?s=60&v=4){kind=small}
There was a problem hiding this comment.
Code Review
This pull request reorganizes and updates the project's agent skills and guidance files for Claude Code, moving the release skill to .claude/skills/create-release/SKILL.md, changing skill triggers from $ to /, and introducing a central CLAUDE.md file. Feedback on these changes identifies two issues in the new release skill file: a nested markdown code block that breaks rendering, and an indented here-document (EOF) that causes a bash syntax error and formatting issues.
Important
The consumer version of Gemini Code Assist on GitHub is being sunset. Starting June 18, 2026, new organization installations will be blocked, and all code review activity will officially cease on July 17, 2026.
For more details on the timeline and next steps, please review the Help Documentation.
There was a problem hiding this comment.
Pull request overview
This PR migrates the repository’s agent/skill configuration and conventions from the prior Codex-style layout to Claude Code conventions, updating skill invocation syntax and relocating the skill definitions accordingly.
Changes:
- Introduces
CLAUDE.mdproject guidance and removes the legacyAGENTS.md. - Moves skills from
.agents/to.claude/and removes the Codex-onlyopenai.yamlmanifest. - Updates skill docs to use slash invocation syntax (e.g.
/new-feature,/feature-flags) and aligns artifact guidance to referenceCLAUDE.md.
Reviewed changes
Copilot reviewed 7 out of 7 changed files in this pull request and generated 1 comment.
Show a summary per file
| File | Description |
|---|---|
| CLAUDE.md | Adds Claude Code project guidance and skill invocation conventions. |
| AGENTS.md | Removes legacy agent guidance file superseded by CLAUDE.md. |
| .claude/skills/new-feature/SKILL.md | Updates invocation wording and artifact instruction references to CLAUDE.md. |
| .claude/skills/feature-flags/SKILL.md | Updates invocation wording and refines guidance language. |
| .claude/skills/create-release/SKILL.md | Adds the Claude Code release skill documentation (needs a Markdown fence fix per comment). |
| .agents/skills/new-feature/agents/openai.yaml | Removes Codex-only manifest. |
| .agents/skills/create-release/SKILL.md | Removes legacy skill location in favor of .claude/ equivalent. |
💡 Add Copilot custom instructions for smarter, more guided reviews. Learn how to get started.
Address reviewer feedback (gemini-code-assist, copilot): - Use a four-backtick outer fence so the nested brew code block renders. - Dedent the release-notes heredoc to column 0 so the <<'EOF' terminates correctly and the notes render without stray indentation. Co-authored-by: Claude <noreply@anthropic.com> Co-authored-by: GitButler <gitbutler@gitbutler.com>
Summary
AGENTS.md→CLAUDE.md,.agents/→.claude/, and deleted the Codex-onlyopenai.yamlmanifest (SKILL.md frontmatter already carries name/description/invocation)./new-feature,/feature-flags,/create-release) and pointed the new-feature artifact instructions atCLAUDE.md.create-release,feature-flags, andnew-featureskills against the Opus 4.8 prompting guidance — literal-scope wording, positive framing, and the reasoning behind each non-obvious rule — preserving every command, path, and fact.Product-level Codex references in
README.mdanddocs/are genuine MCP integration features (the read-only endpoint Codex consumes,codex app-serverresearch) and stay untouched.Test plan
find .claude/skills -type fshows the threeSKILL.mdfiles, noagents//openai.yaml.test -f CLAUDE.md && ! test -e AGENTS.md && ! test -e .agentspasses.grep -rnE '\$new-feature|\$feature-flags|AGENTS\.md|openai|\.agents' CLAUDE.md .claude/skillsreturns nothing.README.md/docs/unchanged in the diff.