Codex Agent Skills package for capability discovery, MCP registry search, skill benchmarking, dynamic tool integration, and safe test-time tool synthesis.

A collection of Skills for Claude Code

A stateful Agent Skill for structured network and system administration labs with durable state, verification, recovery, and handoff workflows.

Agent skills that turn a one-line goal into a dependency-ordered sprint of GitHub issues, then run them one at a time — review-each (HITL) or fully unattended (AFK). No cluster, no Linear: just gh, claude, and your repo.

🌀 Claude Code skills that plan the agent loop, remember what worked, and graduate proven loops into new skills. A loop-design + Reflexion-memory flywheel.

Agent Kaizen skill for Claude Code & Codex: faithful, task-routed Git reference, commands, flags, config keys & refspecs; staging, committing, branching, merging, rebasing, cherry-pick, revert, reset, stash, tags; log/diff/blame/bisect; remotes & push/pull; revision syntax; hooks, submodules, worktrees, reflog recovery.

A pre-submit quality gate for Solana AI Kit skill builders — and a review accelerator for maintainers. Zero dependencies, read-only, 50 tests, SARIF output.

Was sind ChatGPT Agent Skills?

Stand up a rigorous, regression-proof evaluation harness for any LLM/agent system from zero — datasets, scorers, CI gates, and drift monitoring.

Open-source Agent Skill for law-firm AI visibility, GEO/AEO, LLM rankings, competitor share of voice, and evidence-based recommendations.

Claude Code skill that installs an interconnected markdown knowledge base (LLM wiki) into any project.

Two open-source, cross-agent Agent Skills for Claude Code, Codex, and OpenAI: Repo Improvement Architect (bounded repo audits + CI triage) and Avatar Motion Forge (original premium mascot SVG + Lottie/Rive motion).

Claude Code skill: orient yourself in an unfamiliar repo — recon, deduce its purpose, confirm with the user, then ask what to do. Intent-based, language-universal triggering.

An Agent Skill for trustworthy headless verification of AutoCAD & Civil 3D plugins — crash-marker liveness, determinism gating, and an evidence-gated pass/fail checklist.

Quarantine untrusted external text before coding agents act.

Zero-downtime SQL migration safety as an Agent Skill: detect lock hazards, rewrite unsafe DDL, auto-generate rollbacks, gate execution (Postgres-first, wraps squawk + eugene).

Personal Agent Skills for AI coding agents — install and update with gh skill

MCP Server + Agent Skills for decentralised AI-Agent identity — mint DIDs and issue, verify & resolve W3C Verifiable Credentials.

scraper-audit-expert - Claude Code Agent Skill (target LLM: Claude). Part of the mr-bridge.com toolkit.

Some good skills

An open-source arsenal of reusable skills for AI agents — repo understanding, code review, delivery workflows, security pentests, and plug-and-play integrations with Jira, GitHub, GitLab, Jenkins & more. Drop into Claude Code, Codex, or any skill-aware host.

Turn rough ideas into agent-ready prompts with confirmation, safety gates, and Codex/Claude Code dispatch support.

Deploy and manage AI training jobs on rented GPU cloud providers, bare metal servers, and Kubernetes clusters using an automated agent skill.

Modular AI agent skills used across engineering, writing, and research workflows. Model‑agnostic, composable, and production‑tested. Includes engineering, productivity, personal, and experimental skill packs.

Hardcore Audit of Misleading Models, Evidence Reliability, Runtime, and Reality. - Go Get Uncle Matt's HAMMERRR: read-only project-truth audit skill for claim-vs-runtime reality checks.

Production-proven agent skills: agent-swarm, asset-validation, browser-harness, awesome-presentation, workspace-knowledge-graph.

Adversarial critique loop for stress-testing a new idea before you build it - with a filter layer that separates real risks from over-worry. Agent skill for Claude Code / OpenCode.

Use this skill to audit, rewrite, and optimize Airbnb and short-term-rental listings, including titles, descriptions, amenity positioning, photo guidance, guest-fit messaging, and conversion-focused improvement plans.

Reviewed cross-agent Skills, provenance, capability topology, and adaptive Harness governance.

Skills that evolve in public — community agent-skill drops for SkDD

Skills for form-heavy product UI review and field-local feedback.

Make web UIs controllable by browser agents without brittle selectors.

面向中文内容团队的 GEO/AEO Codex Skill，把旧文章、产品页、教程和知识库改造成更容易被 AI 搜索发现、理解、推荐和引用的内容资产

Portable skills, agents, and operating workflows for AI assistants.

Modular operating system for Generative Engine Optimization and AI Search workflows, with skills, rubrics, templates, benchmarks and validation.

Agent Kaizen skill for Claude Code & Codex: Windows PowerShell 5.1 and Visual Studio Developer Shell guidance for any Windows VS Code repo, setup, build & wrapper scripts, native exec (Python, Cargo, npm, MSBuild), MSVC workflows, call operator & argument arrays, $LASTEXITCODE, encoding, logging & validation.

Claude Code skill: a complete STRIDE/LINDDUN threat-modeling engagement — data-flow model, OWASP risk scoring, mitigations, gap loop, and a diffable remediation register. Enterprise AppSec rigor for solo builders.

Claude Code skill: a solo security program as durable artifacts — a living risk register, NIST CSF 2.0 self-assessment, and a lightweight incident-response plan. The right-sized GRC slice, no SOC2 ceremony.

AI agent skills for MCLL Real Estate — luxury residences across Thailand, for sale and rent

Evidence-backed local inventory and safe cleanup for Agent Skills

Agent Kaizen skill for Claude Code & Codex: faithful, identifier-preserving reference for the Discord Developer Platform, bots & apps, slash/application commands, interactions & message components, REST API, Gateway (intents/opcodes/events), resources (channels, guilds, webhooks), OAuth2, permissions, embeds, and monetization.

An honest single-model 'LLM Council' skill for Claude Code: 5 advisors pressure-test a decision, peer-review anonymously, and a chairman flags unanimous agreement as possible shared bias — not corroboration.

mplementation-level engineering-craft practices — testing, failure handling, simplicity, extension boundaries, resource control, API discipline — distilled from gold-standard codebases

高德地图 Agent Skill（无需 MCP）— geocoding/POI/weather/route planning for Claude Code · pi · any shell agent, pure curl. AMap (Gaode) maps, no MCP server.

Plans, ships, and verifies zero-downtime schema + data migrations with an expand/migrate/contract state machine that refuses to skip the unsafe step.

Agent-drivable Remotion pipeline: structured JSON in, a loudness-normalized MP4 out. A production harness that layers on top of remotion-dev/skills.

Scan-gated safety for Agent Skills before they reach Codex, Claude Code, CI, or a registry.

Turn a capability spec into a correct agent/tool architecture: decide skill vs subagent vs MCP vs workflow, design tool contracts, context strategy, safety, and a test harness.

Reusable Claude Code skills that help AI agents collaborate: handoffs, shared task tracking, code-review exchange, ADRs, and changelogs. Installable as a plugin.

Claude skill that counteracts LLM sycophancy: honest assessment over the agreeable one. Mirror-test, steelman-the-opposite, resist caving. Calibrated, not contrarian.

Provider-neutral AI conversation continuity framework for verified session handoffs

A triage-gated process for taking a behavioral skill from works to publish-grade and beyond. Scale the apparatus to the skill; every step leaves a checkable artifact.

AWS Blocks Skills

Metacognitive debugging and Project-Based Learning skill for Codex and Claude AI coding agents

An LLM agent skill for high-performance Entity Framework Core — fast queries, bulk inserts, high-throughput patterns — that fixes code without changing its behavior.

Grok Build skill: read-only second opinions from Claude Code (plan, review, advice)