superpowers
claudecodexCore skills library for Claude Code: TDD, debugging, collaboration patterns, and proven techniques
15 skills · hooks
Progressive Guardrails output style, the write-test skill and its four test agents, the remove-code skill, the guardrail skill, and the node-native skill.
Plugin coding from marketplace ezzabuzaid in ezzabuzaid/agent-plugin. Commands read from the repo's manifest — 3 days ago.
guardrailTurns a rule stated in words ("never use enums", "never git stash", "tests never override env") into a check that a machine runs, or says why no machine can run one. Picks the tool and the rung, confirms that the check exists, is exact and has a trigger, and names its home. Use when a user correctio
node-native-refreshBrings the node-native catalog up to the current Node LTS: rebuilds the surface snapshot, reviews every API that is new or changed since the last review, writes or updates catalog entries with their probes, and re-runs every probe so a gap the LTS closed turns into a mapping. Run inside the agent-pl
node-nativeFinds what the current Node.js LTS already ships that a codebase still does by hand: dependencies with a native replacement (chalk → util.styleText, glob → fs.glob, dotenv → --env-file), hand-written code that re-implements a native API (a deferred() helper → Promise.withResolvers, a JSON clone → st
remove-codeStop owning code the tool already owns. Delete hand-written workarounds, wrappers, monkey-patches, forks, and glue that a dependency provides officially — and reach new goals through a tool's official API or config seam instead of writing code. Applies to additive work, not only deletion: if the cha
write-testDecides which tests are worth writing and how to write them black-box. Gates each test on the break it catches, discovers the repo's real entry point and upstream seams, and checks the gotchas that turn tests into false positives. Use when writing, reviewing, extending, or refactoring tests; when a
test-claim-comparatorRead-only. Checks a write-test claim ledger against the requirement it must cover — the user's request word for word plus the ticket, which it finds and fetches itself — and reports requested behaviors no claim covers and claims weaker than what was asked. Never reads the tests. Launched by the writ
test-discovererRead-only. Maps how a repo's tests must be written before any test is drafted — the runner and what it actually runs, the public surface, the outermost upstream seams, real in-process forms of dependencies, a precedent test worth copying, state and isolation constraints. Launched by the write-test s
test-gotcha-reviewerRead-only adversarial reviewer for test changes. Checks a diff against the write-test skill's Gate and diff-visible gotchas, including production changes made for a test's sake, and reports each hit with file:line and the cheapest way the test could pass while the behavior is broken. Launched by the
test-mutation-proverProves each new test catches the break it claims to. Deliberately breaks the source the way each test says it guards against, in the memory of the test processes (or in a scratch copy, for what Node cannot load), runs that test as the repo's own test target runs it, and reports CAUGHT, SURVIVED, or
Core skills library for Claude Code: TDD, debugging, collaboration patterns, and proven techniques
15 skills · hooks
Up-to-date documentation lookup. Pull version-specific documentation and code examples directly from source repositories into your coding agent's context.
1 MCP server
Build powerful, stateful CLI interfaces for any GUI application using the cli-anything harness methodology.
5 commands
Automates browser interactions for web testing, form filling, screenshots, and data extraction
12 skills
Use Alibaba Open Code Review from local Codex.
2 skills
Claude Code skill for creating and presenting Slidev slidedecks
1 skill
This page updates itself from your repo. Add the badge to your README, or get it Featured to pin it above the organic results with a followed link to your site.