Agent skills for real engineering, not vibe coding. End-to-end loop: PRD → vertical-slice issues → parallel TDD → one consolidated PR. Plus standalone helpers for triage, diagnose, and code-review.
Disciplined diagnosis loop for hard bugs and performance regressions. Reproduce → minimise → hypothesise → instrument → fix → regression-test. Use when user says "diagnose this" / "debug this", reports a bug, says something is broken/throwing/failing, or describes a performance regression.
Create a git branch with the prefix convention (`feature/`, `fix/`, `chore/`, `refactor/`, `env/`) required by GitHub auto-PR workflows. Use when user wants to start a branch, says "new branch", or asks to begin a feature.
Grilling session that challenges your plan against the existing domain model, sharpens terminology, and updates documentation (CONTEXT.md, ADRs) inline as decisions crystallise. Use when user wants to stress-test a plan against their project's language and documented decisions.
Walk a human through the manual-action [HITL] sub-tasks of a parent (PRD) issue — the steps a coding agent physically cannot perform — recording each as an audit trail and marking them done so the dependent [AFK] slices unblock, then hand back to /tdd-parallel. Use when /tdd-parallel skipped HITL slices, when the user wants to clear manual steps before a fanout, or when the user invokes /human-itl directly.
Find deepening opportunities in a codebase, informed by the domain language in CONTEXT.md and the decisions in docs/adr/. Use when the user wants to improve architecture, find refactoring opportunities, consolidate tightly-coupled modules, or make a codebase more testable and AI-navigable.
Own this plugin?
Verify ownership to unlock analytics, metadata editing, and a verified badge. GitHub access is read-only (username + org membership).
Sign in to claimOwn this plugin?
Verify ownership to unlock analytics, metadata editing, and a verified badge. GitHub access is read-only (username + org membership).
Sign in to claimBased on adoption, maintenance, documentation, and repository signals. Not a security audit or endorsement.
Agent skills for real engineering, not vibe coding. Small, composable, model-agnostic — adapt them to your repo instead of letting a process framework take over.
📖 Full guide: superpowers.zsl.dev
[!IMPORTANT] Built for Claude Code. Skills depend on Claude Code's slash commands,
/plugininstall path, and tool surface (Agent,Monitor,Bash,SendMessage,TaskStop). They aren't drop-in for Cursor, Codex, Cline, or general LLM chat. ("Model-agnostic" above means any Claude model — Opus, Sonnet, Haiku — not any harness.) See Compatibility for what porting would look like.
The skills compose into one end-to-end loop. Most days you only touch a few of them.
/setup-zsl-superpowers — configure the issue tracker, triage label vocabulary, domain doc layout, and ship style for this repo. Run once before anything else./grill-me or /grill-with-docs — interview yourself to surface what you're actually building. grill-with-docs also updates CONTEXT.md and ADRs inline./to-prd — synthesise that conversation into a PRD on the issue tracker./to-issues — break the PRD into vertical-slice sub-issues. Children are labeled needs-triage; the PRD parent is auto-relabeled to tracking. Slice titles use the [AFK|HITL] <wave>[<letter>] — <description> format so the dependency graph reads at a glance (same wave = runnable in parallel)./triage — triage each child to ready-for-agent (with an agent brief), ready-for-human, or needs-info. Skip triaging the PRD itself; you just wrote it./human-itl <PRD> — clear all [HITL] slices first: the manual actions a coding agent can't perform (console clicks, credential rotation, sign-off). Records each as an audit-trail comment, marks them done. Must be cleared before /tdd-parallel, not interleaved — the auto-pipeline refuses to start with any [HITL] open. Hard-refuses slices that are really decisions in disguise (those belong upstream in /grill-with-docs + an ADR)./tdd-parallel <PRD> — full-auto pipeline from a PRD to a pushed integration PR. Fans out the unblocked [AFK] ready-for-agent children into parallel /tdd sub-agents in worktrees (/tdd --no-ship). The orchestrator merges every slice branch onto the PRD branch in wave order with --no-ff, runs an integration /code-review --auto (step 4a), auto-chains /verify-coverage --auto (step 4b), and auto-fixes any gaps it finds — gap sub-issues land as ready-for-agent and re-fanout in the next round; loop iterates until gap=0 or a circuit breaker fires (--max-coverage-rounds, default 3). On clean coverage, step 4c delegates the defensive commit to /commit then pushes and opens one consolidated integration PR. Pre-flight (1d) refuses up front if any [HITL] is open or any user story isn't acceptance: automatable — both gates ensure the post-invocation pipeline runs through to PR-push with no human gates in the happy path. PR-style repos only./tdd <child> — single-issue red-green-refactor. Refuses if you point it at a container. On local-markdown trackers, you can also call it with no argument — it scans .scratch/, resolves each open issue's ## Blocked by against the issues/done/ archive, and lets you pick from the unblocked ones./verify-coverage <PRD> — check every PRD ## User Stories entry against the implemented code via tests, not prose. Tier A maps each story to a passing behavioral test; for stories with none, Tier B generates one from the story's observable: tag, mutation-proves it non-vacuous, and runs it. Quarantines failing tests, auto-files gaps as sub-issues of the PRD, and writes a coverage receipt against the verified sha. Almost always chained by /tdd-parallel step 4b in --auto mode (gaps filed as ready-for-agent directly, no triage hop; the orchestrator loops on them). Direct user invocation is for auditing PRDs whose slices shipped via a different path. The human-attestation HITL lane has been removed — non-automatable stories are refused at /to-prd time, so visual/UX/external work goes through a separate manual path that doesn't touch /tdd-parallel.npx claudepluginhub zunosmartlabs/zsl-superpowers --plugin zslUltra-compressed communication mode. Cuts ~75% of tokens while keeping full technical accuracy by speaking like a caveman.
Design fluency for frontend development. 1 skill with 23 commands (/impeccable polish, /impeccable audit, /impeccable critique, etc.) and curated anti-pattern detection.
Behavioral guidelines to reduce common LLM coding mistakes, derived from Andrej Karpathy's observations on LLM coding pitfalls
Frontend design skill for UI/UX implementation
Comprehensive skill pack with 66 specialized skills for full-stack developers: 12 language experts (Python, TypeScript, Go, Rust, C++, Swift, Kotlin, C#, PHP, Java, SQL, JavaScript), 10 backend frameworks, 6 frontend/mobile, plus infrastructure, DevOps, security, and testing. Features progressive disclosure architecture for 50% faster loading.
Access thousands of AI prompts and skills directly in your AI coding assistant. Search prompts, discover skills, save your own, and improve prompts with AI.