Coding & Development

Turn the current conversation into a spec and publish it to the project issue tracker: no interview, just synthesis of what you've already discussed.

Plan a huge chunk of work (more than one agent session can hold) as a shared map of decision tickets on your issue tracker, and resolve them one at a time until the way to the destination is clear.

Generate an interactive bash wizard that walks a human through steps only they can perform. Use when provisioning infrastructure, setting up credentials or CI secrets, walking an unfamiliar third-party dashboard, or running a one-off migration or cutover. Don't invoke this for steps the agent can perform itself.

Implement a specification in code.

Composes an autonomous unattended Claude Code build prompt that survives usage-limit resets via BUILD_LOG.md checkpointing, plus the launcher that restarts it. Use when the user asks for an overnight build, an autonomous build, a build to run while they sleep, an unattended or unsupervised session, or any long build they will not be monitoring. Trigger phrases include "overnight build", "run overnight", "build while I sleep", "autonomous build", "let it run", "unattended". Produces a single copyable prompt; does not run the build.

The user suspects something in your last response was made up, or contradicts

Convergent, action-first problem decomposition (the terse twin of /fanout). Use ONLY when the user types /breakdown or says "break this down" / "break it down" / "decompose this". Takes ONE tangled or ambiguous problem and returns the single smallest next ACTION plus a collapsed full plan. Terse, no hand-holding, no capability narration. Do NOT use for open-ended ideation with many valid answers (that is /fanout), and do NOT use when the task is already one clear action.

Runs a ranked, evidence-backed audit of a product, codebase, or subsystem, ending in a verify list that separates proven findings from inferred ones. Use when the user asks for an audit, a gap analysis, a deep read, a bug hunt, a readiness check, a re-audit, a state reconciliation, or asks what they are not seeing. Trigger phrases include "audit", "gap analysis", "what am I missing", "deep audit", "readiness", "bug hunt", "review the whole thing". Produces findings and a verify list; does not fix.

Decision and migration guide for Prisma ORM MongoDB projects on v6, which have no upgrade path to v7. Use when a MongoDB project asks about upgrading Prisma, when "upgrade to prisma 7" comes up in a project with provider = "mongodb", or when evaluating a move to Prisma Next. Triggers on "upgrade prisma mongodb", "prisma 7 mongodb", "mongodb prisma migration", "prisma next mongodb".

Verification gate for a Next.js-on-Vercel + Supabase + GitHub Actions stack, to run before claiming anything is shipped, deployed, live, merged, passing, or done. Use before saying a change is in production, before merging a PR, before reporting CI green, and before writing any status into a document — even if the user does not ask for verification. Trigger phrases include "is it live", "did it deploy", "ship it", "merge it", "is CI green", "confirm prod", "did that work". Encodes traps where the obvious check returns a false pass.

Prisma Compute deployment and hosting guide. Use whenever the user mentions Prisma Compute, `prisma.compute.ts`, `defineComputeConfig`, deploying or hosting a Prisma app, `@prisma/cli app deploy`, `compute:deploy`, `create-prisma --deploy`, `PRISMA_SERVICE_TOKEN`, Compute auth/workspaces, apps/deployments/build logs/domains, localhost vs `0.0.0.0`, deploy port binding, or framework deploy readiness for Hono, Elysia, Next.js, TanStack Start, Astro, Nuxt, Svelte, Nest, Turborepo, or custom/prebuilt artifacts.

Turns an ambitious request into a self-judging, self-improving agent run instead of a first draft. Sets a concrete external quality bar, splits the goal into judgeable pieces, gives each one a separate critic with clean context, forces blind A/B against the bar, and loops with no fixed round count. Routes every spawned agent to the cheapest model and effort that can do its job, so fan-out does not burn tokens. Use for ANY complex or open-ended build, not only games. Trigger when asked to build, create, design, refactor, redesign, optimize or "make excellent" any artifact (app, site, UI, game, API, service, pipeline, report, article, campaign, research), and on goal, loop, gauntlet, fan out, subagents, AAA, production quality, best in class, keep improving, compare against a reference, or any concern about token cost, budget, model choice or reasoning effort in a multi-agent run. Also when a broad goal arrives with no implementation detail, or when about to ship a first draft as final.

Use when the user asks to run Codex CLI (codex exec, codex resume) or references OpenAI Codex for code analysis, refactoring, or automated editing

Instruments code so production behavior is visible and diagnosable. Use when adding logging, metrics, tracing, or alerting. Use when shipping any feature that runs in production and you need evidence it works. Use when production issues are reported but you can't tell what happened from the available data.

Default contract for new static HTML pages from screenshots, bitmap mockups, or raster designs. Use for HTML projects, landing pages, clone a screenshot, tokenize CSS, TUNE blocks, theme tokens, studio-web repair after clone, layout packs. Do not use for Wix Velo-only edits or when the user names a different pipeline.

Establishes a project's quality bar as a written contract and stops agents quietly lowering it. Interviews the user on which dimensions matter, supplies sane default thresholds when they have no number in mind, records everything in CONSTRAINTS.md, and watches the diff for a weakened bar — new @ts-ignore or eslint-disable suppressions, skipped or deleted tests, assertions stripped out, unimplemented stubs, thresholds edited down. Use when no quality bar is written down, when the user says "set up constraints" or "define our standards", when the user wants dimensions they care about — accessibility, web performance, coverage — set up as enforced constraints, when an agent keeps silencing checks or skipping tests to get to green, when you need a coverage or performance threshold and don't know what number to pick, or when an agent writes more code than anyone will read.

Use when user asks to 'lint agent configs', 'validate skills', 'check CLAUDE.md', 'validate hooks', 'lint MCP'. Validates agent configuration files against 456 rules.

Harden a plan with independent Claude/Codex review, then optionally build and cross-inspect it. Start in either Claude Code or Codex: the host plans and the other provider reviews. Use for claudex this plan, claudex-loop, or the legacy crucible trigger; not for trivial edits.

Use at the START of work in an unfamiliar or half-remembered codebase, or when the user says "understand this codebase", "get up to speed", "where do I start", "what is this repo", "/begin", or asks what they were last working on. Derives the important and complex parts from the import graph, git churn and recent PRs, links every hot file to the user-facing surface it serves, and writes BEGIN.md — a living knowledge graph of the code that stays honest on every push.

Independent adversarial review and deterministic release gating for code changes, using a multi-model reviewer panel over OpenRouter (or any OpenAI-compatible router, or an MCP transport like Composio). Use whenever the user asks for an adversarial review, red-team review, multi-model review, independent review, release gate, pre-merge or pre-release verification, "is this safe to ship/merge/deploy", or asks to verify a change with models other than the one that wrote it. Also use when the user asks whether work is "done" on production software and wants proof rather than assurance.

Use BEFORE any multi-agent fan-out, swarm, or Workflow orchestration — and when assigning a model tier to any delegated unit of work. Confidence-gated model routing: rubric-based base-tier assignment (cheap→standard→frontier→apex), objective escalation triggers (verification failure, disagreement, uncertainty flags), hysteresis rules, and the structured handoff payload that lets higher tiers resolve only the residue lower tiers couldn''t. Triggers on "fan out", "swarm", "parallel agents", "which model", "assign tiers", "dispatch", "model routing", "token cost".

Codex Skill for cleaning Chinese film and TV public-opinion samples, clustering viewpoints, and generating a verifiable offline workbench

agentmujo-skills-hub

Recently updated agent-skill-related GitHub repository: yutaoshao/agent-skills.