Skills

A set of skills for Kotlin, Jetpack Compose, Android development, and grounded writing.

With the skills CLI:

npx skills add chrisbanes/skills

Or install as a Claude Code plugin:

/plugin marketplace add chrisbanes/skills
/plugin install chrisbanes-skills@chrisbanes-skills

Or install as a Codex plugin:

codex plugin marketplace add chrisbanes/skills --ref main
codex plugin add chrisbanes-skills@chrisbanes-skills

Or install as an OpenCode plugin:

{
  "plugin": ["chrisbanes-skills@git+https://github.com/chrisbanes/skills.git"]
}

See .opencode/INSTALL.md for details.

• Working on Compose state or effects? Start with compose-state-and-effects.
• Investigating recomposition, stability, or jank? Start with compose-performance.
• Reviewing Flow or coroutine architecture? Start with kotlin-concurrency-and-flow.

• using-chrisbanes-skills — route Kotlin and Jetpack Compose work to the focused skills; current Claude Code versions also activate it when working with .kt or .kts files.

• compose-state-and-effects — decide state ownership and effect lifecycle for local UI state, screen state holders, Flow collection, callbacks, cleanup, navigation, snackbar, analytics, and focus requests.

• compose-performance — diagnose stability, deferred reads, composition contracts, and cross-phase back-writing from concrete runtime evidence.

• compose-component-design — design caller-placeable Compose APIs whose variable visual regions are caller-provided slots.
• compose-animations — choose Compose animation APIs for visibility, value targets, coordinated transitions, and content swaps; align with official quick guide and decision tree.
• compose-focus-navigation — design and test keyboard, TV, D-pad, and focus-first Compose navigation behavior.

• compose-ui-testing-patterns — choose between plain UI tests, semantics assertions, key/focus tests, interaction state tests with MutableInteractionSource, screenshot tests, and integration tests.

• kotlin-concurrency-and-flow — review coroutine ownership, cancellation, Flow state/event modeling, sharing, replay, and one-shot delivery.
• kotlin-control-flow — write and review Kotlin branching with subject when, guard conditions, sealed exhaustiveness, smart casts, nullable branching, and early returns.
• kotlin-api-design — choose function owners, semantic domain types, and Kotlin Multiplatform platform boundaries.

• grounded-writing — draft or revise clear, evidence-led writing of any length, including review comments and replies, without inventing personal claims.

• gradle-run — run every agent-initiated Gradle command through a compact-output wrapper; Gradle-centered workflows use one read-only diagnostic owner while parents retain edits.
• implement-with-subagents — implement or review supplied-task orchestration through separate implementation owners, preserving atomic work, task-scoped acceptance, repair ownership, and the installed implement dependency.
• to-plan — create a repository-aware implementation plan from one ready GitHub issue or an in-chat task, with a provider-neutral implementation handoff.
• run-github-project — set up, review, or operate the repository's GitHub Project workflow; preserve live authority, human Planning work, unknown outcomes, epics, checkpoints, triage, and authorized execution boundaries.
• shepherd — autonomously poll open PRs and MRs, triage review comments, and switch CI failures into a full local verification-and-repair cycle.

This is a breaking taxonomy change. Replace the removed entrypoints as follows:

Skills live at skills/<skill-name>/SKILL.md, flat (no language nesting). The name: in the SKILL.md frontmatter must match the directory name.

Frontmatter is validated against skills.schema.json — name and description are required, name must be kebab-case. The router also uses Claude Code's optional paths extension. Clients that do not support this extension must ignore the paths field rather than rejecting the skill.

Release versions use SemVer-compatible CalVer: YYYY.M.D without zero-padded month or day values, for example 2026.6.17.

Keep .claude-plugin/plugin.json, .codex-plugin/plugin.json, and new Git release tags on the same version. Existing zero-padded tags from before this policy map to the non-padded manifest version, so 2026.06.16 maps to 2026.6.16. Only bump versions when publishing an installable release.

To publish a release, run the Release workflow from GitHub Actions. Leave the version input empty to use today's UTC YYYY.M.D version, or provide a specific non-zero-padded CalVer value. Use the dry-run option to validate without creating a commit, tag, or GitHub release.

Before pushing, lint skills (frontmatter schema + markdown):

npm install
npm run lint

This also runs on CI for all PRs.

For a taxonomy change, also run the durable cluster behavior evaluation. It checks routing, required references, safeguards, exceptions, and finish gates at the public agent-facing seam.

The advisory evaluator tests concrete scenarios modelled on real-world coding work, with expected outcomes and no-change controls. It compares no-skill, forced-skill, and automatic-routing runs. Baseline is the no-skill result, automatic is the headline result, and restraint checks that a skill does not make an unnecessary change. Scorecards also compare subject-side tokens, tool calls, completed turns, elapsed time, and total attempted work per successful outcome. The table reports the latest available result for each skill and correctness metric. These scores were produced using gpt-5.6-terra with medium reasoning, judged by gpt-5.6-sol with high reasoning. Results are model- and reasoning-specific; other configurations may perform differently. These are not merge or release gates. See evals/README.md for evaluation setup and reproducibility.

Values are per-run medians, baseline → automatic, followed by the automatic percentage change. These subject-only measurements use the latest complete, same-run evidence available for each suite and include failed runs and negative controls. Multi-skill scenarios contribute to every targeted skill row. A turn is one completed Codex turn; time remains environment-sensitive. The source runs, selection rules, and detailed scorecards are in the evaluation change record.

Apache 2.0