Loading...
Loading...
Found 671 Skills
Multi-agent discussion rooms — debate or poll a problem from multiple perspectives. Standalone or invoked by other skills as a sub-routine. Mode=debate: N agents argue in rounds, converge. Mode=poll: N agents independently analyze, aggregate by consensus. Not for implementation (use system-architecture). Not for verification (use review-chain). For clarifying requirements first, see discover. For decomposing work after a decision, see task-breakdown.
Use when the user asks about chaos engineering, fault injection, resilience testing, or HA verification for a SPECIFIC AWS service (e.g., RDS, EKS, MSK, ElastiCache, DynamoDB, S3, Lambda, OpenSearch, etc.). Triggers on "chaos testing on [service]", "fault injection for [service]", "how to test HA of [service]", "FIS scenarios/actions for [service]", "[service] failover testing", "[service] resilience testing", "[service] 混沌测试", "[service] 故障注入", "[service] 高可用验证", "对 [service] 做混沌实验", "test my [service]", "verify my [service] is resilient". Use this skill even when the user phrases it casually like "test my RDS" or "how resilient is my MSK cluster".
Use when you need framework-agnostic WireMock guidance — stub design, JSON or programmatic mappings, precise request matching, response bodies and faults, classpath fixtures, isolation and reset between tests, verification of calls, dynamic ports and base URLs, and avoiding flaky stubs — without choosing Spring Boot, Quarkus, or Micronaut. Part of the skills-for-java project
Converts Markdown files to one MS Word document per file using plugin-local scripts. V2 includes L5 Delegated Constraint Verification for strict binary artifact linting.
Executes full-project QA like a real user by discovering the repository verification and E2E contracts, running build, lint, test, and startup commands, exercising core workflows end-to-end through CLI, HTTP, and browser interfaces, requiring automated regression coverage for supported critical flows, fixing root-cause regressions, and rerunning the full gate. Uses the agent-browser companion skill for Web UI validation when a web surface exists. Use when validating a branch, release candidate, migration, refactor, or risky commit. Do not use for static code review only, one-off unit test edits, planning test cases, or architecture brainstorming without execution — use qa-report for planning and documentation.
End-to-end Claude Design handoff to pull request: imports a handoff bundle from claude.ai/design, generates Storybook stories and Playwright tests, runs diff-aware browser verification, and opens a PR with the bundle URL, before/after screenshots, and coverage delta embedded in the body. The one-shot 'design URL in, reviewable PR out' workflow. Use when a designer or PM hands you a Claude Design URL and you want a PR back without intermediate steps.
Zero-context verification that every number, comparison, and scope claim in the paper matches raw result files. Uses a fresh cross-model reviewer with NO prior context to prevent confirmation bias. Use when user says "审查论文数据", "check paper claims", "verify numbers", "论文数字核对", or before submission to ensure paper-to-evidence fidelity.
Ultra-lightweight channel for refactor processes - used when changes are obviously too small to justify the full scan → design → apply three-stage workflow. AI directly identifies 1-3 low-risk optimization points, confirms with the user once, modifies in-place using classic methods, and validates itself by running tests. No scan checklist, no design documentation, no multi-step HUMAN verification required. Trigger scenarios: When the user says "quick refactor", "small refactor", "simply optimize XX function", "modify directly", "skip all those steps", and the scope of changes is clearly limited to a single function/single component, with tests available for self-validation.
Reliable end-to-end engineering workflow for debugging, root-cause analysis, minimal patching, and verification in production codebases. Use when Codex needs to investigate a failure systematically, trace execution, test hypotheses, implement a correct fix, validate the resolution, and check for regressions before declaring the task complete.
Background knowledge for droid-control workflows -- not invoked directly. Deliverable verification against commitments.
Launch multiple sub-agents in parallel to execute tasks across files or targets with intelligent model selection, quality-focused prompting, and meta-judge → LLM-as-a-judge verification
Fireflies.ai platform help — AI meeting note-taker with GraphQL API, webhooks (V1 + V2), AskFred AI, real-time events, and Fred bot that joins Zoom/Meet/Teams to transcribe. Use when Fireflies transcripts not syncing to CRM, webhooks not firing or signatures failing HMAC verification, hitting 50 req/day or 60 req/min rate limits on the GraphQL API, building a transcript pipeline from Fireflies to Snowflake/BigQuery/warehouse, migrating from Webhooks V1 to V2, the Fireflies bot not joining calls or users wanting to disable auto-join, deciding between Free, Pro ($10), Business ($19), or Enterprise ($39) tier, or wiring AskFred or Real-time API into an internal app. Do NOT use for comparing Fireflies vs Fathom/Avoma/Gong or selecting a note-taker (use /sales-note-taker) or reviewing a single sales call for coaching (use /sales-call-review).