智能体技能是一个包含指令、脚本和资源的文件夹，智能体按需加载。与 MCP 服务器不同，它不运行自己的进程，因此在上下文上更省，也更容易审查。
第 265–288 条，共 935 条
smykla-skalski
Use when the user wants a refactoring review of code, an app, a module, or a diff. Scans the target for code smells and git hotspots, reviews it through seven opposed refactoring lenses (Fowler, Uncle Bob, Feathers, Beck, Metz, Ousterhout, Tornhill), synthesizes a safety-first sequenced refactoring plan, then runs a separate adversarial agent that red-teams the plan before returning it. Triggers o
smykla-skalski
Audit and fix CLAUDE.md files using a tiered binary checklist based on official Anthropic best practices and community guidelines. Use when the user asks to "review CLAUDE.md", "audit CLAUDE.md", "score CLAUDE.md", "improve CLAUDE.md", or "fix CLAUDE.md".
smykla-skalski
Diagnose and fix flaky e2e tests and general connectivity issues in service mesh environments (Kuma, Istio, Linkerd, Consul). Trigger when: a user mentions intermittent test failures, "test is flaky", e2e failures in CI that don't reproduce locally, Ginkgo/Gomega test files that fail sometimes, 503/connection refused errors, mTLS handshake failures, pods not getting traffic, xDS NACKs or warming r
smykla-skalski
Staff-engineer-level code review that goes beyond correctness to evaluate architectural alignment, system-level implications, failure modes, performance, scalability, backward compatibility, observability, security, and cross-team impact. Use when reviewing a PR (URL or diff), analyzing code changes for architectural fitness, or when the user asks for a thorough/staff-level/senior review of code c
zstevovich
MANDATORY after every successful APD pipeline commit on Codex — when apd:apd_pipeline_state shows next_step='commit' and a commit was made. Verify tests from a clean state, show the pipeline report, present four options (push, push+PR, keep local, discard), execute only after the user picks. Triggers on "after commit", "push", "PR", "pull request", "deploy", "next", "pipeline done", "shipped", com
zstevovich
Use BEFORE writing the APD spec-card.md and calling apd:apd_advance_pipeline('spec', ...) on Codex whenever the task is vague, broad, ambiguous, or has multiple reasonable interpretations. Ask ONE question at a time, present 2-3 approaches when there are real choices, converge on a design the user explicitly approves. Works hand in hand with `.apd/rules/brainstorm.md` when that file exists. Trigge
zstevovich
MANDATORY before re-dispatching the builder after any APD verifier block or critical reviewer finding on Codex. Use on any bug, test failure, build failure, error, crash, regression, or unexpected behavior. Four-phase systematic debugging — root cause investigation, pattern analysis, hypothesis + controlled test, fix with TDD. Escalate after three failed hypotheses instead of piling on more guesse
ARCASSystems
The front door to the Pipeline pack: turn a name into a tracked, worked deal. Trigger on "turn this name into a deal", "track this prospect", "help me with my pipeline", "I met someone, help me pursue them", "work this lead", "new business with <name>", "follow up with <name>", or any open-ended business-development request where the user has a person or company but no clear next step. You name wh
diegocconsolini
Process Word documents (1MB-50MB+) with token optimization through local extraction and caching. Preserves formatting, tables, and document structure for policy analysis, contract review, and technical reports.
diegocconsolini
Navigate Claude Code documentation and get quick answers about configuration, MCP, plugins, and best practices.
diegocconsolini
Generate standards-compliant SBOMs (CycloneDX 1.6 + SPDX 2.3) for a project across stacks (npm, Python, Go, Rust, Java, containers), with optional quality scoring, vulnerability scan, VEX, and signing. Adds zero dependencies to the target project. Use when the user asks for a software bill of materials, dependency inventory, supply-chain manifest, or CycloneDX/SPDX output.
diegocconsolini
Use this skill to install ready-made Claude Code hooks for security — a ConfigChange compliance audit trail (logs every settings/skill change), plus PreToolUse guards and optional HTTP notifications for security-relevant events. Covers command, http, and prompt hook types.
diegocconsolini
Use when applying the same security fix across many files or repos at once — large-scale vulnerability remediation. Provides 7 copy-paste /batch recipes (XSS, HTTPS, SQLi, input validation, log PII, secret rotation, CVE patches) each with a safety tier and false-positive trap, plus scan_diff.py to prove the migration introduced no new HIGH/CRITICAL findings.
openshift-eng
Label a GitHub PR for squash merging by posting a /label tide/merge-method-squash comment
openshift-eng
A simple greeting command that demonstrates plugin structure
quay
Understand a Quay bug report from JIRA or GitHub. Fetches ticket details, checks for existing work, classifies as UI vs backend, and proposes an investigation plan. No code changes.
quay
Orchestrates the Quay bug-fix workflow through 9 phases with confidence-based gating. Reads confidence from phase artifacts to advance automatically, post JIRA comments, or escalate to the user.
smykla-skalski
Daily AI news digest covering technical advances, business news, and engineering impact. Aggregates from research papers, tech blogs, HN, newsletters. Use daily for staying current on AI developments.
smykla-skalski
End-to-end ship a GitHub issue — implement, adversarial code review via subagent, adversarial manual testing via subagent, open PR, wait for Copilot review + green CI, address feedback, merge, close issue. Use when given a GitHub issue URL and asked to implement and ship it.
smykla-skalski
Run the council workflow from a normal Copilot session only when the user explicitly asks for council review, multi-persona critique, debate, design review, code review, architecture feedback, UX review, or tradeoff analysis. Do not use it for commit, stage, merge, approval, or generic pre-commit requests. Accept the same mode syntax as the bundled council reviewers: `core|auto|core-eng|core-ux|co
diegocconsolini
Use this skill to scan a Claude Code plugin (or any npm/PyPI package tree) for security issues before installation. Detects dangerous functions, code obfuscation, hardcoded credentials, schema problems, and known 2024-2026 supply-chain attack IOCs (Shai-Hulud, Nx s1ngularity, chalk/debug clipper, Axios RAT, litellm/.pth), mapped to MITRE ATT&CK/ATLAS, CVE, and OWASP.
mnzralee
A generator-plus-critic loop that improves output against an explicit rubric before acceptance. Use when you have clear acceptance criteria and iteration measurably helps (specs, prompts, important docs, tricky implementations). Includes a seed eval set for regression-testing prompt and rule changes.
vaisala-xweather
This skill should be used when working with the Xweather MapsGL Android SDK (mapsgl-android-sdk / com.xweather.mapsgl) - setting up MapboxMapController, adding or removing weather layers via LayerCode or WeatherService configs, styling with StyleValue and Expression, custom sources and layers, legends, data inspector presentations, timeline animation, layer masks, or integrating the AAR/JitPack de
vaisala-xweather
This skill should be used when working with the Xweather MapsGL SDK for Apple platforms (the MapsGL iOS/iPadOS/macCatalyst/visionOS SDK) — installing it via Swift Package Manager, CocoaPods, Carthage or xcframeworks, creating a MapboxMapController or MapLibreMapController, and adding, removing, styling, animating or inspecting MapsGL weather layers in Swift. Use it whenever a task mentions MapsGL
