agent-skills

Reviews and guides LLM/AI application engineering — prompts, caching, RAG, agent loops and tools, resilience, model migration, evals, prompt-injection defence, and observability. Invoke with /ai-engineering for "review my prompt", "set up RAG", or "write evals".

Authors performant CSS-first web animations and picks the right feedback for an interaction — transitions, View Transitions, scroll-driven motion, Motion for React, reduced motion, and perceived performance. Triggers on "animate this", "hover effect", "feel faster", "skeleton loader", "/animations".

Authors performant React Native / Expo animations and gestures with Reanimated and react-native-gesture-handler on the UI thread; web animations go to `animations`. Triggers on "reanimated", "react native animation", "gesture handler", "swipe gesture", "/animations-native".

Uploads local screenshots and recordings to GitHub and embeds them in a PR or issue body or comment as inline-rendering user-attachments URLs, via `gh --attach` with a curl fallback; never writes placeholders. Triggers on "attach screenshots to the PR", "upload this recording to the PR", "/attach-media".

The phase machinery (0–7) behind the `aw` dispatcher: task intake through a tested draft PR in an isolated worktree, gated by the plan.md brief and the run-checks.mjs sentinel, with companions that never block. Not the entry point — natural-language requests for autonomous work go to the `aw` skill; invoke /autonomous-workflow only to run the phases directly.

Ships coding work to a tested draft PR from one opt-in entry point: detects the tier, runs Micro/Lite here, and on Full writes a short brief with sealed checks, works to them here, and converges once with a local review loop before the PR; prints a paste-ready /goal. Use when asked to "implement autonomously", "independently", "in isolation", "in a worktree", "end-to-end", "all the way to a PR", "ship this", "land this", "take care of this", "handle this without me", or /aw.

Writes the plan artifact in `.agent/{branch}/` — `plan.md` as a brief (task, guardrails, definition of done, turn cap, file changes) plus `checks.yaml`, its executable acceptance checks, sealed for `run-checks.mjs`. Use after planning and on every plan iteration; triggers on create plan, write plan artifact, regenerate plan, iterate on plan.

Writes `.agent/{branch}/walkthrough.md`, a one-page post-hoc review of finished work — what changed, what was assumed, and the evidence — for the reviewer to read after the fact. Use at Phase 6 before the draft PR; triggers on create walkthrough, generate walkthrough, write walkthrough artifact.

One-time (but safely re-runnable) setup flow that scaffolds a project's aw-tester aw-target: detects auth strategy, captures storage state, writes .claude/aw-targets/<target>.yml, and validates with a smoke spec. `--target local` (default) scaffolds the local dev target; `--target preview` scaffolds the PR-preview target ui-verify runs against (and is what `/ui-verify setup` delegates to). Also records the repo's UI surface (what counts as a UI change) so the is-ui-diff gate is accurate per repo, and the preview auth profile. Re-runs detect the existing aw-target and only re-prompt for what broke or changed. Triggers on "/aw-setup", "setup aw-tester", "scaffold aw-target".

Runs aw-tester UI specs in-session through the claude-in-chrome extension against an already-logged-in Chrome, emitting the same verdict block; falls back to aw-tester when the extension is absent. Triggers on "run the spec in chrome", "verify with the chrome driver", or `ui-verify run --driver chrome`.

Proves a feature works end to end, locally or on the PR preview: reuses the brief and specs `/aw` wrote, gets the environment ready, probes it in the browser (ui-verify) with concurrency and fault cases, then reads what the services recorded (telemetry-verify), fixes what it finds on a local stack one commit per finding, and reports with real screenshots. `setup` writes the committed `.agents/aw/` config every suite skill reads. Triggers on "/aw-verify".

Applies behavioral science (COM-B, B=MAP, EAST, defaults) to change behavior in teams and organizations (employees, managers, or engineers adopting a practice or tool) and in product flows; plans rollouts, workshops, and presentations; never recommends a dark pattern or sludge. Triggers on "how do I get my team to X", "why don't people do X", "roll this out", "nudge", "/behavioral-design".

Generates a personal markdown changelog of merged or closed pull requests authored by the current user and Linear tickets the user closed or worked on, over a configurable window (default 7 days), grouped by feature area (e.g. Dashboards, Agent0). Inputs sourced from `gh search prs --author=@me` and the Linear MCP. Use for weekly recaps, status updates, performance reviews, or end-of-sprint summaries. Triggers on "changelog", "what have I done", "weekly summary", "my recent work", "recap my week", "/changelog".

Selects the chart type and visualization library for React/Next.js and Expo/React Native data visualization, with accessibility and anti-pattern guardrails. Triggers on "what chart should I use", "visualize this data", "build a dashboard", "pick a chart library", "/charting".

Diagnoses and fixes failing GitHub Actions CI, verifies locally, pushes, and confirms the checks pass on the new revision — never weakens checks. Triggers on "CI is failing", "fix the CI", "the build is red", "/ci-auto-fix".

Language-agnostic skill for authoring and reviewing code with low cognitive complexity, readability, and long-term maintainability. Covers guard clauses, single-responsibility functions, type-driven design (illegal states unrepresentable, branded primitives, discriminated unions), schema-first validation, single source of truth for union metadata, functional core + imperative shell, idempotency, and neighbour-pattern symmetry. Pairs with `tdd` for new code (rules apply in GREEN/REFACTOR). Stack-specific extensions live under `rules/stacks/<stack>/` (React, Next.js today; drop a subdirectory for any other language or framework). Use during PR review, after writing new code, in TDD GREEN/REFACTOR, or when asked to "improve quality", "make this readable", "reduce complexity", "deduplicate", "clean this up", or "/code-quality".

Rates confidence that work fully solves the stated requirement, in plan, code (default), or analysis mode; plan mode adds deterministic rule checks, and a failed rule caps the gate at 89% regardless of LLM score. Use before autonomous execution, after implementation, or during investigation; triggers on "confidence check", "validate plan", "rate confidence", "quality gate", "/confidence".

Writes a short narrative PR description (≤ 25 lines, hard ceiling 40), converges the branch with review-loop in local mode, pushes, opens a draft PR, runs review-loop until every thread is resolved, and hands red CI to ci-auto-fix. Flags: --split (2–4 dependency-ordered draft PRs), --quick, --no-review, --no-simplify, --no-quality, --no-pre-review, --no-post-review, --no-ui-verify, --no-media; invoke with /create-pr.

Scaffolds, diagnoses, and prunes agent skills. Modes: `scaffold` (default — runs the job once without a skill on a real input, records every correction, then writes a skill made of only those corrections), `diagnose` (retrospective failure analysis that emits a confidence-gated unified diff against any skill declaring a diagnostic surface), `prune` (deletes unused skills and trims or deletes those whose with/without-skill eval shows no lift over a same-length dummy). Triggers on "create a skill", "scaffold a skill", "new SKILL.md", "diagnose this skill", "why did the skill miss this", "prune my skills", "clean up my skills", "/create-skill".

Adversarial pre-mortem of a plan, code change, or bug diagnosis — cited findings over a fixed failure taxonomy plus a mandatory steelman; never scores or fixes, one pass per run. Modes plan (default), code, analysis; `deep` runs 3–5 parallel persona lenses; triggers on "critical", "challenge this", "pre-mortem", "red-team this", "deep critical review", "review from every angle", "/critical".