feat(commands): compile conventional command prompts into Cursor and Claude artifacts (#100 stage 2)

[ScriptedAlchemy]

Summary

• Adds the commands/*.md authoring surface for Compile capability-aware host component bundles #100 stage 2: Markdown prompts with optional closed frontmatter (description, argumentHint, allowedTools, model, disableModelInvocation, plus the peeled bundle-only targets), fresh AB4920–AB4926 diagnostics, duplicate-name and capability-aware target validation.
• Lowers per host, never leaking authoring-only fields: Cursor emits the plain prompt body (its documented commands surface is frontmatter-free Markdown) with a commands: './commands/' manifest pointer; Claude regenerates documented kebab-case frontmatter (allowed-tools, argument-hint, description, disable-model-invocation, model) from validated fields; Codex and portable record honest unavailable capability rows, and the composite bundle emits Claude-format commands while deliberately omitting the Cursor pointer (documented in the generated AGENTS.md) because pinned Cursor evidence does not establish tolerance for Claude frontmatter.
• Applies the feat(rules): lower conventional Cursor rules with honest capability rows (#100 stage 2) #207 review-fix patterns from the start: command source/provenance.sourcePath are canonicalized in canonicalizeNormalizedModel (root-independent modelDigest, regression-tested across two temp roots), and commands are wired into inspection skip accounting (kind: 'command').
• Capability tables gain dated evidence (Cursor plugin schema commands pointer; Claude plugins-reference commands contract); Claude adapter revision 1.4.0, Cursor 1.4.0, with recomputed capability shas; artifact-validator enforces the .md-only commands layout; packaging boundary tests keep command-free projects byte-identical.

Test plan

• pnpm typecheck
• Scoped unit suites: command-config, rule-config, cursor-adapter, plugin-bundle, host-adapters, adapter-capability-states, adapter-metadata, artifact-validator (43 passed)
• Integration: api.test.ts including digest root-independence and inspection skip accounting (31 passed)
• pnpm lint clean

[changeset-bot]

🦋 Changeset detected

Latest commit: 35c2a8b

The changes in this PR will be included in the next version bump.

This PR includes changesets to release 1 package

Name	Type
agent-bundle	Minor

Not sure what this means? Click here to learn what changesets are.

Click here if you're a maintainer who wants to add another changeset to this PR

[chatgpt-codex-connector]

Codex Review Summary

This comment shows the latest Codex review activity on this pull request.

Review	Status	Commit	Review trigger
📝 Code Review	✅ Completed 2026-09-01T21:46:20.101864Z	35c2a8b	PR opened

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" or "@codex security review".

Codex reacts with 👀 while any review is running, comments if it has suggestions, and reacts with 👍 once all reviews finish with no findings.

[pkg-pr-new]

npm i https://pkg.pr.new/ScriptedAlchemy/agent-bundle@219

npm i https://pkg.pr.new/ScriptedAlchemy/agent-bundle/create-agent-bundle@219

npm i https://pkg.pr.new/ScriptedAlchemy/agent-bundle/@agent-bundle/runtime@219

commit: 35c2a8b

[chatgpt-codex-connector]

💡 Codex Review

Here are some automated review suggestions for this pull request.

Reviewed commit: 35c2a8bcf7

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".