feat: package AgentGuard as Claude Code + Codex skills#443
Conversation
- Add skills/agentguard/SKILL.md (Anthropic Claude Skills format) - Add skills/codex/agentguard.md (Codex skill format) - Add skills/README.md orienting readers to both formats Both skill files link back to root SKILL.md as the canonical reference to avoid content drift. PyPI install path stays agentguard47; import path stays from agentguard. Awesome-agent-skills directory submission deferred per their "real community usage / not brand-new" gate. Will queue follow-up in ~2 weeks once skill files have install evidence.
![chatgpt-codex-connector[bot]](https://avatars.githubusercontent.com/in/1144995?s=60&v=4){kind=small}
There was a problem hiding this comment.
💡 Codex Review
Here are some automated review suggestions for this pull request.
Reviewed commit: f33c5dd0b8
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".
There was a problem hiding this comment.
Pull request overview
This PR adds packaged “skill” entrypoints for AgentGuard so it can be installed/used in standardized locations for Claude Code (Anthropic skills format) and Codex (Codex skill format), plus a small skills/ README to route readers to the right file. It also updates the accompanying planning/research/QA artifacts to reflect this documentation-only task.
Changes:
- Add a Claude Code–style skill doc at
skills/agentguard/SKILL.md. - Add a Codex-format mirror at
skills/codex/agentguard.mdand a shortskills/README.mdindex. - Update
WORK_PLAN.md,RESEARCH.md, andQA_REPORT.mdto match the new scope.
Reviewed changes
Copilot reviewed 6 out of 6 changed files in this pull request and generated 6 comments.
Show a summary per file
| File | Description |
|---|---|
| WORK_PLAN.md | Re-scopes the work plan toward skill packaging docs. |
| skills/README.md | Adds a small index + install snippet for the skills directory. |
| skills/codex/agentguard.md | Adds Codex skill entrypoint doc and quickstart snippet. |
| skills/agentguard/SKILL.md | Adds Claude Code / Anthropic skill entrypoint doc. |
| RESEARCH.md | Captures format/submission research and task notes. |
| QA_REPORT.md | Records QA checks for the doc-only change set. |
|
@bmdhodl this PR has been open 3+ days; review or close |
Summary
skills/agentguard/SKILL.md(Anthropic Claude Skills format) andskills/codex/agentguard.md(Codex format) so AgentGuard can be installed as a packaged skill in either harness.skills/README.mdorienting readers to which format to use.SKILL.mdto prevent content drift.Why
Distribution: VoltAgent's awesome-agent-skills directory + the broader Claude Code / Codex skill ecosystem expects skills to live in standardized paths. Root
SKILL.mdexists but is not in the conventionalskills/<name>/layout. This PR makes AgentGuard installable as a drop-in skill.Awesome-agent-skills submission
Deferred intentionally. Their CONTRIBUTING.md gate is "real community usage and proven adoption (not brand-new submissions)." A same-day submission fails that gate. Plan: ship these skill files, let them mature ~2 weeks, then submit with download/issue evidence. Follow-up will be queued separately.
Test plan
agentguard47and import pathfrom agentguard import ...match rootSKILL.mdgithub.com/bmdhodl/agent47,pypi.org/project/agentguard47)Artifacts
WORK_PLAN.md,RESEARCH.md,QA_REPORT.mdupdated for this taskRisk
Low. Documentation-only diff. No code, no deps, no auth, no schema, no workflows.