Repository navigation
docs(plugins): state how a skill names another plugin's skill - #345
Merged
Merged
Conversation
Each plugin installs alone, but the instruction files gave no rule for an instruction that applies a skill of another plugin. Skills settled it one at a time, and the reason babysit-pr stops without the engineering plugin while create-pr continues was written nowhere. The policy names the two forms, soft requirement and hard requirement, and chooses between them by whether a safe fallback exists.
…ilable The comment rules applied writing:technical-writing with no condition, so a session without the writing plugin had an instruction it could not follow.
Codex Review Summary: ✅ CompletedThis comment shows the latest Codex review activity on this pull request.
ℹ️ About Codex in GitHubYour team has set up Codex to review pull requests in this repo. Reviews are triggered when you
Codex reacts with 👀 while any review is running, comments if it has suggestions, and reacts with 👍 once all reviews finish with no findings. |
There was a problem hiding this comment.
🟢 Approval recommended
The reviewed documentation and aligned plugin version updates have no remaining active issues identified.
No open issues 🎉
What changed in this PR
Documents how skills should depend on skills from separately installed plugins, distinguishing soft requirements with fallbacks from hard requirements that must be checked before side effects. It also updates the engineering plugin version.
Changes:
- Adds cross-plugin soft- and hard-requirement guidance to
plugins/AGENTS.md. - Updates the engineering comments reference to conditionally use the technical-writing skill.
- Bumps the engineering plugin version to
3.3.1in both manifests.
| File | Description |
|---|---|
| plugins/engineering/references/COMMENTS.md | Updates the comments guidance to conditionally reference the technical-writing skill. |
| plugins/engineering/plugin.json | Bumps the portable engineering plugin version to 3.3.1. |
| plugins/engineering/.claude-plugin/plugin.json | Keeps the Claude extension manifest aligned at version 3.3.1. |
| plugins/AGENTS.md | Defines soft and hard requirements for skills that depend on skills from other plugins. |
Review effort: Lite
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
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.
Each plugin installs alone, but no rule said how a skill should name a skill of another plugin. Each
skill settled the question by itself. The reason
git:babysit-prstops without theengineeringplugin while
git:create-prcontinues was not written down, and a plannedgit:shipskill needsthat rule before it takes a requirement on
engineering.This change adds a "Requirements between plugins" section to
plugins/AGENTS.md. An instructionthat applies a skill of another plugin takes one of two forms:
available, and the work continues without it. When the report would otherwise imply the skipped
work happened, the applying skill states the fallback and reports the omission.
side effect, and stops with the plugin to install when it is absent. One check covers every skill
of that plugin.
The choice depends on the fallback, not on invocation policy. A hard requirement applies only when,
without the other skill, the applying skill would take a side effect outside the checkout on a
decision the other skill supplies. A skill applied only to wording always has a safe fallback. The
rule could not depend on invocation policy, because
babysit-pris implicitly invokable and itshard requirement is justified.
Existing references
I checked every instruction in
plugins/that applies another plugin's skill against the policy.One did not comply:
engineering/references/COMMENTS.mdappliedwriting:technical-writingwith nocondition. It now applies the skill when that skill is available, and
engineeringmoves to 3.3.1.The two terms are defined in the policy section and are not added to the ubiquitous language
section, so each has one definition.