chore(skills): mirror model/enrichment/hosted-key/council skills into .agents/skills and expand add-model touchpoints

[waleedlatif1]

Summary

• Added .agents/skills/ mirrors (SKILL.md + agents/openai.yaml) for add-model, validate-model, add-enrichment, and add-hosted-key, and a SKILL.md for council — these existed only as .claude/commands/ and lacked the cross-tool skill format the other integration skills have. Every command now has a skill mirror.
• Expanded add-model with the repo-side touchpoints that aren't data-driven: hosted-key billing (a model under openai/anthropic/google is auto-served with Sim's rotating key and billed via getHostedModels()/shouldBillModelUsage()), tests with hardcoded model IDs that bun run lint won't catch, net-new provider-code handling, and the embedding/rerank/new-provider escape hatches.
• Added a matching hosting/billing audit check to validate-model.

Type of Change

• Chore (tooling/docs)

Testing

Verified every .claude/commands/* now has a .agents/skills/* mirror, frontmatter matches the existing skill format, and the enriched touchpoint content carried into both the command and the skill. Biome ignores these paths (markdown/yaml), no code changed.

Checklist

• Code follows project style guidelines
• Self-reviewed my changes
• Tests added/updated and passing
• No new warnings introduced
• I confirm that I have read and agree to the terms outlined in the Contributor License Agreement (CLA)

[vercel]

The latest updates on your projects. Learn more about Vercel for GitHub.

Project	Deployment	Actions	Updated (UTC)
docs	Ready	Preview, Comment	Jun 4, 2026 7:16pm

[cursor]

PR Summary

Low Risk
Documentation-only changes to agent skills and Claude commands; no application code, auth, or billing logic is modified.

Overview
This PR adds agent skill mirrors under .agents/skills/ for workflows that previously lived only as .claude/commands/: add-enrichment, add-hosted-key, add-model, validate-model, plus council (SKILL only, no openai.yaml by design). Each integration skill ships SKILL.md frontmatter and an agents/openai.yaml card so cross-tool discovery matches existing skills like add-block.

add-model guidance is expanded in both the new skill and .claude/commands/add-model.md: a new step for non–data-driven repo touchpoints—getHostedModels() / shouldBillModelUsage() auto-billing for openai/anthropic/google, provider tests lint won’t run, net-new provider code, and embedding/rerank/new-provider escape hatches—with a hosted/billing row in the verification report and extra anti-patterns.

validate-model is aligned in the skill and .claude/commands/validate-model.md: canonical URL table points at both command and skill mirrors, plus a Hosting / billing audit checklist for models under hosted providers.

No runtime or apps/sim application code changes—documentation and agent tooling only.

^{Reviewed by Cursor Bugbot for commit 324b596. Configure here.}

[greptile-apps]

Greptile Summary

This PR creates .agents/skills/ mirrors for four existing .claude/commands/ — add-model, validate-model, add-enrichment, and add-hosted-key — giving each a SKILL.md plus an agents/openai.yaml agent card. It also adds a council skill (SKILL.md only, intentionally no yaml). Alongside the new mirrors, add-model is expanded with repo-side touchpoints not covered by its data-driven model entry (hosted-key auto-billing via getHostedModels(), hardcoded model IDs in tests, net-new provider-code handling, and embedding/rerank escape hatches), and validate-model picks up a matching hosting/billing audit checklist item and a disambiguated cross-reference to add-model.

• New skill mirrors: five skills under .agents/skills/ created from scratch, content parity with their .claude/commands/ counterparts confirmed.
• add-model expansion (both command and skill): new Step 4 covering hosted billing enrollment, test assertions, and net-new API behavior; anti-pattern list extended accordingly.
• validate-model update (both command and skill): hosting/billing checklist added; ambiguous add-model.md path reference replaced with fully-qualified dual-path reference.

Confidence Score: 5/5

No code changes — purely documentation and tooling files; safe to merge.

All changes are markdown and YAML files under .agents/skills/ and .claude/commands/. No TypeScript, no migrations, no runtime logic is touched. The new skill mirrors faithfully reflect their updated command counterparts, previous review comments have been addressed, and the intentional omission of council's agent card is now explicitly documented in its frontmatter.

No files require special attention.

Important Files Changed

Filename	Overview
.agents/skills/add-model/SKILL.md	New skill mirror of .claude/commands/add-model.md; includes all expanded Step 4 repo-side touchpoints (hosted billing, test assertions, provider code) and updated anti-pattern list. Content matches the updated command file.
.agents/skills/validate-model/SKILL.md	New skill mirror of .claude/commands/validate-model.md; hosting/billing checklist section added, cross-reference to add-model disambiguated with dual paths. Content in sync with the updated command file.
.agents/skills/add-enrichment/SKILL.md	New skill mirror for add-enrichment; covers the full provider-cascade workflow, hosted-key verification gate, and client-safety constraints. No issues found.
.agents/skills/add-hosted-key/SKILL.md	New skill mirror for add-hosted-key; covers BYOK registration, pricing research, hosting config, UI wiring, and key-pooling analysis. No issues found.
.agents/skills/council/SKILL.md	New skill for the council meta-utility; intentionally ships no agents/openai.yaml, now documented explicitly in frontmatter to prevent future confusion.
.claude/commands/add-model.md	Expanded with new Step 4 covering hosted-key auto-billing, test assertions with hardcoded model IDs, net-new provider API behavior, and wrong-family escape hatches; steps renumbered and anti-patterns extended. Matches the new skill mirror.
.claude/commands/validate-model.md	Step 2 cross-reference disambiguated; hosting/billing checklist section added. Kept in sync with the new validate-model skill mirror.

Flowchart

%%{init: {'theme': 'neutral'}}%%
flowchart TD
    A[".claude/commands/"] --> B["add-model.md\n(updated)"]
    A --> C["validate-model.md\n(updated)"]
    A --> D["add-enrichment.md\n(existing)"]
    A --> E["add-hosted-key.md\n(existing)"]

    F[".agents/skills/"] --> G["add-model/\nSKILL.md + openai.yaml"]
    F --> H["validate-model/\nSKILL.md + openai.yaml"]
    F --> I["add-enrichment/\nSKILL.md + openai.yaml"]
    F --> J["add-hosted-key/\nSKILL.md + openai.yaml"]
    F --> K["council/\nSKILL.md only\n(no yaml by design)"]

    B <-->|"mirror"| G
    C <-->|"mirror"| H
    D <-->|"mirror"| I
    E <-->|"mirror"| J

    B --> L["New Step 4: Repo touchpoints"]
    C --> M["New checklist: Hosting/billing"]

Loading

_{Reviews (2): Last reviewed commit: "chore(skills): document council yaml omi..." | Re-trigger Greptile}

[waleedlatif1]

@greptile

[waleedlatif1]

@cursor review

[cursor]

✅ Bugbot reviewed your changes and found no new issues!

Comment @cursor review or bugbot run to trigger another review on this PR

^{Reviewed by Cursor Bugbot for commit 324b596. Configure here.}