agentsclimarketplace

Context & AI engineering

25,745 rows, by stacks then stars

1,865 of these skill files have been read, by 1,328 distinct authors, and what they tell an agent counted. What Context & AI engineering authors agree on, and what they forbid

  • Session handoff

    bhaumikmaan/claude-code-master-skills/skills/session-handoff Skill

    3★ repo

    Create structured session notes for continuity across conversations. Produces fixed-template notes with immutable section headers, session titles, and token-budgeted content. Use when saving session state, creating handoff notes, preparing for context compaction, or when asked to summarize work for continuity.

  • Task orchestration

    bhaumikmaan/claude-code-master-skills/skills/task-orchestration Skill

    3★ repo

    Coordinate complex multi-step tasks across parallel workers. Handles research, synthesis, implementation, and verification phases. Use when a task is too large for a single pass, requires parallel investigation, involves multiple files or subsystems, or needs coordinated implementation and verification.

  • Brainstorm

    momentmaker/kaijutsu/skills/core/brainstorm Skill

    3★ repo

    Multi-agent ideation against a free-form prompt. Wraps `jutsu swarm brainstorm "<prompt>"` to run claude / codex / antigravity in parallel with three different angles (claude=long-horizon framing, codex=code-pattern grounding, antigravity=cross-domain analogy), synthesizes into a ranked list of approaches with tradeoffs. Use when the user says "brainstorm", "ideas for", "how should we approach", "options for", or invokes /brainstorm. Designed to feed into `decide` (for ADR capture of the picked option) or `refactor-plan` (for execution of the picked option).

  • Doc review

    momentmaker/kaijutsu/skills/core/doc-review Skill

    3★ repo

    Universal multi-agent QA gate for markdown artifacts (specs, plans, decision records, RFCs, design docs). Wraps `jutsu swarm doc-review <path>` to run claude / codex / antigravity in parallel with tailored lenses (claude=completeness, codex=implementability, antigravity=consistency), synthesizes into a single review with a disagreement table. Designed to be the final-review pass for any artifact-producing skill (spec-driven-development, planning-and-task-breakdown). Use when the user says "review this spec", "review this plan", "check this RFC", or invokes /doc-review. Prints to stdout; cache + replay supported.

  • Dream

    momentmaker/kaijutsu/skills/core/dream Skill

    3★ repo

    Pre-implementation interrogation. Walks any topic through 8 cognitive lenses to surface hidden assumptions, weak premises, orthogonal angles, temporal decay, and adversarial misuse BEFORE the topic becomes a spec / plan / commit. Use when the user says "what do you think about X?", "should we do X?", "is X worth pursuing?", "let's brainstorm" (note: this is NOT brainstorm — see "vs brainstorm preset" below), or invokes /dream. Composes via jutsu swarm dream for multi-agent matrix mode.

  • Pr review

    momentmaker/kaijutsu/skills/core/pr-review Skill

    3★ repo

    Adversarial multi-agent pull-request review. Runs claude / codex / antigravity in parallel via `jutsu swarm pr-review`, each with a tailored prompt (claude=architecture+correctness, codex=edge cases, antigravity=cross-file patterns), then synthesizes into a single markdown review with a disagreement table. --full mode adds a Pass-2 round-robin debate where agents critique each other. --strict adds a lie-to-them filter on the synthesis draft. Posts the result as a PR comment when --post-comment is set; edits prior kaijutsu-pr-review comments in place. Use when the user says "review this PR", "code review", "review the diff", "look at PR #N", or invokes /pr-review. Refuses to send the diff unless `.kaijutsu/pr-review.yaml` has `allow-multi-model: true` (first-run prompt persists this). Hard-blocks on a pre-flight secrets scan unless --allow-secrets is passed.

  • Refactor plan

    momentmaker/kaijutsu/skills/core/refactor-plan Skill

    3★ repo

    Multi-agent refactor planning — files + goal in, ordered step plan with risk per step out. Wraps `jutsu swarm refactor-plan <path>... --goal "<goal>"` to run claude / codex / antigravity with three angles (claude=architectural decomposition, codex=stepwise risk, antigravity=pattern consistency). Designed to feed into `planning-and-task-breakdown` for execution. Use when the user says "plan this refactor", "how do I restructure", "step plan to extract X", or invokes /refactor-plan.

  • Security audit

    momentmaker/kaijutsu/skills/core/security-audit Skill

    3★ repo

    Multi-agent security audit. Runs claude / codex / antigravity in parallel through three CVSS-aligned lenses (claude=auth+data flow, codex=injection+priv-esc, antigravity=dep+supply-chain). Wraps `jutsu swarm security-audit --pr <n>` (diff mode) or `jutsu swarm security-audit <path>...` (files mode). Output is a threat model with attack vectors + prioritized mitigations. --full mode defaults ON because security findings are high-stakes; the Pass-2 debate filters false-positives that would erode trust. Use when the user says "audit this for security", "threat model", "check for vulns", "review the auth flow", or invokes /security-audit.

  • Clerk cli

    arthjean/skills/skills/clerk-cli Skill

    no license3★ repo

    Operate Clerk from a coding-agent terminal with the official clerk CLI and bundled Backend API helpers. Use when the agent needs to inspect or manage Clerk applications, users, organizations, memberships, invitations, sessions, impersonation, JWT templates, instance configuration, allowlists, blocklists, domains, OAuth applications, SAML connections, webhooks, or Clerk tokens; invoke clerk-cli or the Clerk CLI; diagnose Clerk authentication or project linking; call the Clerk Backend API; or translate a legacy Clerk MCP management workflow to shell commands. Do not use for application runtime integration through Clerk SDKs, frontend sign-in flows, continuous monitoring, or Clerk MCP server configuration.

  • Context7 mcp

    arthjean/skills/skills/context7-mcp Skill

    no license3★ repo

    This skill should be used when the user asks about libraries, frameworks, API references, or needs code examples. Activates for setup questions, code generation involving libraries, or mentions of specific frameworks like React, Vue, Next.js, Prisma, Supabase, etc.

    The same file is in 3 other repositories

  • Handoff

    arthjean/skills/skills/handoff Skill

    no license3★ repo

    Compact the current conversation into a handoff document for another agent to pick up.

    The same file is in 11 other repositories

  • Mcp server dev

    arthjean/skills/skills/mcp-server-dev Skill

    no license3★ repo

    Build MCP (Model Context Protocol) servers in TypeScript with @modelcontextprotocol/sdk. Use when writing, reviewing, or refactoring MCP server code: (1) Creating MCP servers with McpServer, (2) Registering tools with registerTool, inputSchema, outputSchema, Zod validation, (3) Defining resources and resource templates, (4) Defining prompts with arguments, (5) Transports: StdioServerTransport, NodeStreamableHTTPServerTransport, SSE, (6) Tool annotations (readOnlyHint, destructiveHint, idempotentHint), (7) Error handling and isError responses, (8) Dynamic tool loading and tool list change notifications, (9) Middleware patterns for MCP tools, (10) Testing MCP servers with vitest, (11) Publishing and configuring for Claude Code, Cursor, Windsurf, (12) Any @modelcontextprotocol/sdk imports.

  • Neon cli

    arthjean/skills/skills/neon-cli Skill

    no license3★ repo

    Operate Neon Postgres from a coding-agent terminal with the official neonctl CLI and psql, without relying on the Neon MCP server. Covers projects, local project linking, branches, databases, roles, connection strings, SQL execution, transactions, schema diffs, migrations, EXPLAIN, slow-query inspection, and Management API gaps. Use when the agent needs to inspect, query, or modify a Neon database; create, restore, reset, diff, or delete branches; run migrations; manage Neon projects or roles; or when the user says neon-cli, neonctl, query my Neon DB, create a Neon branch, or list my Neon projects. Do not use for application runtime integration with Neon, ORM schema design, continuous monitoring, or Neon MCP server configuration.

  • Posthog cli

    arthjean/skills/skills/posthog-cli Skill

    no license3★ repo

    Operate PostHog from a coding-agent terminal with the official @posthog/cli agent API and bundled REST helpers. Use when the agent needs to inspect, query, create, update, or delete PostHog insights, dashboards, feature flags, experiments, cohorts, persons, events, surveys, recordings, error-tracking issues, warehouse resources, CDP functions, LLM analytics, organizations, or projects; run HogQL or typed analytics queries; upload source maps or debug symbols; translate a PostHog MCP workflow; or when the user says posthog-cli, PostHog CLI, PostHog API, HogQL, query my PostHog data, or manage my PostHog flags. Do not use for application runtime instrumentation through PostHog SDKs, continuous polling, PostHog MCP server configuration, or dashboard-only billing and instance administration.

  • Resend cli

    arthjean/skills/skills/resend-cli Skill

    no license3★ repo

    Operate Resend from a coding-agent terminal with the official resend-cli and bundled REST helpers. Use when the agent needs to send, schedule, inspect, forward, or cancel emails; manage inbound email, domains, API keys, contacts, contact properties, segments, topics, broadcasts, templates, webhooks, automations, events, suppression entries, or API logs; diagnose Resend authentication or delivery; invoke resend-cli; or translate a Resend MCP workflow to shell commands. Do not use for application runtime integration through a Resend SDK, continuous monitoring, non-Resend email providers, or Resend MCP server configuration.

  • Skill creator

    arthjean/skills/skills/skill-creator Skill

    no license3★ repo

    Interactive guide for creating new skills (or updating existing skills) that extend Claude's capabilities with specialized knowledge, workflows, or tool integrations. Walks the user through use case definition, frontmatter generation, instruction writing, and validation. Use when user asks to create a new skill, build a skill, update an existing skill, says "skill-creator", "create a skill for", "build a skill that", or wants help designing skill structure and content. Do NOT use for general prompting questions or non-skill configuration tasks.

  • Skill doctor

    arthjean/skills/skills/skill-doctor Skill

    no license3★ repo

    Audit a Claude Code skill or workflow for structural quality, idempotency, and best-practice compliance. Runs 30+ binary checks grounded in Anthropic's official skill authoring docs, Agent Patterns research, and production patterns. Produces a severity-weighted report (FAIL/WARN/NOTE). Use when the user says 'audit this skill', 'check my skill', 'skill-doctor', 'review my workflow', or provides a path to a skill directory. Do NOT use for codebase audits (use meta-audit) or PRD reviews (use meta-review-prd).

  • Vercel cli

    arthjean/skills/skills/vercel-cli Skill

    no license3★ repo

    Operate Vercel from a coding-agent terminal with the official Vercel CLI and bundled REST helpers. Use when the agent needs to inspect or manage Vercel teams, projects, deployments, build or runtime logs, environment variables, domains, aliases, Edge Config stores, webhooks, deployment-protection bypass tokens, log drains, promotions, rollbacks, or redeployments; when the user asks to deploy to Vercel, diagnose a Vercel deployment, check Vercel logs, manage Vercel environment variables, or invoke vercel-cli; or when translating a Vercel MCP workflow to shell commands. Do not use for authoring application code that merely runs on Vercel, application runtime integration through @vercel/sdk, continuous monitoring, or Vercel MCP server configuration.

  • Create agent skill en

    roebi/agent-skills/skills/create-agent-skill-en Skill

    no license3★ repo

    Creates a new Agent Skill that follows the agentskills.io open specification. Use this skill whenever a user or agent wants to author a new skill, scaffold a skill directory, write a SKILL.md, add reference files, or add scripts to an existing skill. Activate for phrases like: "create a skill", "write a skill", "new skill for X", "scaffold a skill", "add a skill to my repo", or "make a SKILL.md".

  • Terminal cli en

    roebi/agent-skills/skills/terminal-cli-en Skill

    no license3★ repo

    Reference for operating in a Linux terminal. Use when the user or agent needs to run bash commands, combine commands with && or pipes, read files, sort and deduplicate output, write and execute shell scripts, work with the filesystem (files, links, mounts, permissions), install or update packages, get help on commands, manage containers with Podman, make HTTP requests with curl, manage Kubernetes clusters with kubectl, manage Helm charts, handle environment variables, or launch AI coding agents (aider-chat, claude). Activate this skill for any task involving: bash, shell, terminal, CLI, script, pipe, podman, kubectl, helm, curl, apt, pip, environment variable, aider, or any Unix command.

  • Distill sessions

    leek/agent-skills/skills/distill-sessions Skill

    3★ repo

    Personal collection of agent skills for Claude

  • Compaction handoff

    uzysjung/uzys-agent-harness/.claude/skills/compaction-handoff Skill

    3★ repo

    Create a compact, reconstructible checkpoint immediately before context compaction. Persist only durable facts and decisions, overwrite one current resume anchor, verify Git and open-PR state, and emit one concrete next action plus a short /compact line. Use for "컴팩션 준비해줘", "컴팩션하고 이어서 진행할 수 있게 준비해줘", "핸드오프 준비해줘", or "prepare for compaction", or equivalent. Repeated execution must be idempotent and must not grow state files without bound.

  • Eval harness

    uzysjung/uzys-agent-harness/.claude/skills/eval-harness Skill

    3★ repo

    Formal evaluation framework for Claude Code sessions implementing eval-driven development (EDD) principles

  • Model orchestration

    uzysjung/uzys-agent-harness/.claude/skills/model-orchestration Skill

    3★ repo

    Apply the fixed model-role and thinking-effort policy whenever work is delegated to subagents or a model/effort choice is made: the orchestrator (top-tier model) DIRECTLY sets service direction, reviews plan/spec documents (with multi-persona-review), improves shipped features, and hunts performance/security problems; plan/spec/planning documents are AUTHORED by Opus, which also owns core implementation and V&V, at xhigh or above; repetitive implementation and E2E tests go to Sonnet at high or above — never delegate below those effort floors. Use whenever you are about to spawn an Agent/Task/Workflow worker, pick a model for a subtask, set a thinking/effort level, assign verification, or hand off orchestration because the current model's quota is exhausted. Trigger on "위임해", "에이전트로 돌려", "오케스트레이션", "모델 역할분담", "어떤 모델로", "effort 얼마로", "thinking level", "서브에이전트", or in English "delegate this", "spawn an agent for", "which model should", "route this task", "verify with", "orchestrate". Fire even when the user doesn't name the policy — any delegation decision is in scope.

  • Agent introspection debugging

    uzysjung/uzys-agent-harness/templates/skills/agent-introspection-debugging Skill

    3★ repo

    Structured self-debugging workflow for AI agent failures using capture, diagnosis, contained recovery, and introspection reports.

    The same file is in 5 other repositories

  • Compaction handoff

    uzysjung/uzys-agent-harness/templates/skills/compaction-handoff Skill

    3★ repo

    Create a compact, reconstructible checkpoint immediately before context compaction. Persist only durable facts and decisions, overwrite one current resume anchor, verify Git and open-PR state, and emit one concrete next action plus a short /compact line. Use for "컴팩션 준비해줘", "컴팩션하고 이어서 진행할 수 있게 준비해줘", "핸드오프 준비해줘", or "prepare for compaction", or equivalent. Repeated execution must be idempotent and must not grow state files without bound.

  • Eval harness

    uzysjung/uzys-agent-harness/templates/skills/eval-harness Skill

    3★ repo

    Formal evaluation framework for Claude Code sessions implementing eval-driven development (EDD) principles

  • Model orchestration

    uzysjung/uzys-agent-harness/templates/skills/model-orchestration Skill

    3★ repo

    Apply the fixed model-role and thinking-effort policy whenever work is delegated to subagents or a model/effort choice is made: the orchestrator (top-tier model) DIRECTLY sets service direction, reviews plan/spec documents (with multi-persona-review), improves shipped features, and hunts performance/security problems; plan/spec/planning documents are AUTHORED by Opus, which also owns core implementation and V&V, at xhigh or above; repetitive implementation and E2E tests go to Sonnet at high or above — never delegate below those effort floors. Use whenever you are about to spawn an Agent/Task/Workflow worker, pick a model for a subtask, set a thinking/effort level, assign verification, or hand off orchestration because the current model's quota is exhausted. Trigger on "위임해", "에이전트로 돌려", "오케스트레이션", "모델 역할분담", "어떤 모델로", "effort 얼마로", "thinking level", "서브에이전트", or in English "delegate this", "spawn an agent for", "which model should", "route this task", "verify with", "orchestrate". Fire even when the user doesn't name the policy — any delegation decision is in scope.

  • Claude sessions

    gerodp/hermes-productivity-skills/development/claude-sessions Skill

    3★ repo

    List your Claude Code sessions for any time window — start, duration, project, message counts, title, models, token usage, API-equivalent cost, and an A–D cost-efficiency grade with per-session recommendations, straight from the local transcripts.

  • Data pipeline architect

    JayRHa/AgentSkills/data-pipeline-architect Skill

    3★ repo

    Designs robust ETL/ELT data pipelines covering ingestion, idempotency, schema evolution, orchestration, and data quality validation. Use this skill when the user asks to design, build, or review a data pipeline, ingest data from APIs/databases/files into a warehouse or lake, set up batch or streaming ETL/ELT, choose an orchestrator (Airflow, Dagster, Prefect, dbt), make a pipeline idempotent or backfill-safe, handle late-arriving or duplicate data, manage schema drift/evolution, add data quality or freshness checks, or model incremental/CDC loads.

  • Prompt engineer

    JayRHa/AgentSkills/prompt-engineer Skill

    3★ repo

    Designs, refines, and systematically evaluates LLM prompts using structure, role framing, few-shot examples, explicit output contracts, and reasoning scaffolds. Use this skill when the user wants to write or improve a prompt, build a system prompt, craft few-shot examples, reduce hallucination or refusals, enforce a JSON/structured output, design an LLM-as-judge or eval rubric, debug inconsistent or low-quality model outputs, or compare prompt variants ("optimize this prompt", "why is the model ignoring my instructions", "make it return valid JSON", "write a prompt that...", "evaluate these prompts").

  • Rag pipeline designer

    JayRHa/AgentSkills/rag-pipeline-designer Skill

    3★ repo

    Designs end-to-end retrieval-augmented generation (RAG) systems by making principled choices for chunking, embeddings, vector indexing, retrieval, reranking, prompt assembly, and offline evaluation. Use this skill when the user wants to build or improve a RAG pipeline, asks about chunking strategy, embedding model selection, hybrid search, reranking, "my retrieval is bad / hallucinating / missing context", chunk size tuning, vector database choice, or how to measure retrieval quality (recall@k, MRR, nDCG, faithfulness).

  • Herdr pair

    hcaiano/skills/skills/herdr-pair Skill

    3★ repo

    Persistent Claude and Codex coworking inside one Herdr tab. Use when the user asks for Herdr pairing, another workflow needs a live peer, or input begins with `[agent` or includes `[herdr-pair control`. For inbound traffic, including after context compaction, run receive and reply through the bundled sender instead of local output.

  • Fetcher llm streaming

    Ahoo-Wang/skills/plugins/ahoo-fetcher-skills/skills/fetcher-llm-streaming Skill

    3★ repo

    A central aggregation repository for Agent Skills from Ahoo-Wang's open source projects.

    The same file is in 1 other repository

  • Fetcher llm streaming

    Ahoo-Wang/skills/sources/fetcher/skills/fetcher-llm-streaming Skill

    3★ repo

    Use when consuming Server-Sent Events, LLM token streams, OpenAI-style streaming chat responses, Fetcher eventstream helpers, response prototype extensions, stream termination detection, result extractors, or ReadableStream async iteration.

  • Relay handoff stop

    liutao773680119-cmyk/relay-dev-skills/skills/relay-handoff-stop Skill

    3★ repo

    Use when ending a session before switching tools/models, and a complete handoff snapshot must be persisted. 中文常见触发:做一次标准交接、按接力开发规则收尾、切换前先交接。

  • Relay handoff stop

    liutao773680119-cmyk/relay-dev-skills/starter/relay-kit-v1/global-skills/relay-handoff-stop Skill

    3★ repo

    Use when ending a session before switching tools/models, and a complete handoff snapshot must be persisted. 中文常见触发:做一次标准交接、按接力开发规则收尾、切换前先交接。

  • Trackbed orchestrate

    Orfi/trackbed/copilot/skills/trackbed-orchestrate Skill

    3★ repo

    Internal Trackbed skill (invoked by trackbed, not by the user). Drive the ongoing orchestration loop for a Jira epic or a standalone project — read the manifest, compute the next unblocked phase, show the rail view, hand each phase to an executor (Copilot CLI itself, or a Copilot custom agent), record progress and per-phase notes, and absorb runtime roadmap changes. Use whenever a roadmap already exists and work needs to advance, resume, or be re-checked. The roadmap is the single source of truth — always re-read it.

  • Pev external executor

    myksyut/pev-harness/skills/pev-external-executor Skill

    3★ repo

    外部 LLM CLI (v3.5.0 では OpenAI Codex) を Execute phase の実装エンジンとして subprocess invoke する skill。 plan.md (Mode A) または task description + triage.json + cwd context (Mode B) と team-conventions を 1 prompt にまとめて codex exec --json --output-schema に渡し、 codex が workspace-write sandbox 内で file を編集する。 codex 不在 / 未認証 / timeout / non-zero exit 時は executor agent に fallback signal を返す (= Claude native 実装に degrade)。 v2.0 pev-external-reviewer の対称形 (reviewer ではなく executor 版)。

  • Pev external reviewer

    myksyut/pev-harness/skills/pev-external-reviewer Skill

    3★ repo

    外部 LLM CLI (v2.0 では OpenAI Codex) を Reviewer として subprocess invoke する skill。 git diff + plan.md + rubric を組み立てて codex exec --json --output-schema に渡し、 structured JSON output を verify.json に merge する。 timeout / non-zero exit / 認証失敗時は自動で dual-claude にfallback (graceful degrade)。

  • Pev linear sync

    myksyut/pev-harness/skills/pev-linear-sync Skill

    3★ repo

    Linear MCP server 経由で plan.md / verify.json を Linear Issue と双方向 sync。inbound (Linear → spec 抽出)、issue-first (実装前に issue 作成 + branch checkout、v3.3.0+)、outbound success (Done + コメント)、outbound fail (failure summary コメント) の4方向。Linear MCP plugin (`@plugin_linear_linear`) が install済みで認証済みであることが前提

  • Pev subagent memory

    myksyut/pev-harness/skills/pev-subagent-memory Skill

    3★ repo

    ~/.claude/pev/{task_id}/ 配下にsubagent (主にexecutor) の memory directoryを標準化。並列実行時の情報共有と次セッションへの引き継ぎを実現

  • Skill creator

    myksyut/pev-harness/skills/skill-creator Skill

    3★ repo

    Create new skills, modify and improve existing skills, and measure skill performance. Use when users want to create a skill from scratch, edit, or optimize an existing skill, run evals to test a skill, benchmark skill performance with variance analysis, or optimize a skill's description for better triggering accuracy.

    The same file is in 66 other repositories

  • Aicrew handoff

    AKSoftCode/aicrew/codex-skills/aicrew-handoff Skill

    3★ repo

    Generate a compact cross-tool handoff block from current session state for switching tools or models.

  • Eval capture

    GhostlyGawd/recursive-harness/skills/eval-capture Skill

    3★ repo

    Snapshot a just-completed task into the regression corpus (evals/corpus/). Use after any task where the user accepted the result AND future harness changes could plausibly regress it — recurring task shapes, tasks that needed corrections before acceptance, anything involving taste the user articulated. Also use when /meta-retro reports a category with misses but no eval coverage. Without corpus growth, "harness improvement" is unmeasurable drift.

  • Loop prompt architect

    GhostlyGawd/recursive-harness/skills/loop-prompt-architect Skill

    3★ repo

    Consultant procedure for authoring self-contained loop prompts. Trigger when the user asks for a prompt that runs "until it's done/verified/bug-free", wants /goal + /loop templates, says "loop Claude until X is finished", needs a fuzzy goal turned into checkable done-criteria, or returns with a failed loop-run to autopsy. Forges wishes into 3-7 binary Definition-of-Done criteria, per-iteration verification, a state artifact, a stuck protocol, and three exits (success/budget/blocked). AUTHORS loop prompts as deliverables — NOT the built-in /loop runner, NOT build-loop.

  • Routing learnings

    GhostlyGawd/recursive-harness/skills/routing-learnings Skill

    3★ repo

    The anti-auto-memory router. Use this EVERY time you notice something worth remembering — a repeated mistake, a user correction, a procedure that worked, a preference, a project quirk — and especially whenever you feel the urge to "just note this down" in CLAUDE.md or memory. Also use during /retro to classify each mined learning. Routing prose into the wrong artifact is the primary failure mode of this harness; when in doubt, trigger this skill.

  • Worktree

    GhostlyGawd/recursive-harness/skills/worktree Skill

    3★ repo

    Create + manage git worktrees in this harness — when to isolate vs when one is overhead, how to EnterWorktree/ExitWorktree (they live at .claude/worktrees/<name>), and the gotchas that bite here: results don't auto-merge to main; state/ is gitignored so `harness` writes in a worktree miss the main ledger; committed memory/ rides in; shared DB/ports aren't isolated; the guard protects the trunk, not a worktree's own copies. Use whenever a second session opens, independent tasks fan out, or file-mutating agents run beside other work — the user never asks for a worktree or recites a gotcha.

  • Use agent docs

    fmind/agent-docs/skills/use-agent-docs Skill

    3★ repo

    Before web-searching or fetching docs for a project topic, check `.agents/docs/INDEX.md` and the matching `DOC.md` for a navigation map of curated upstream URLs. No-op if `INDEX.md` is missing.

  • Figma handoff

    m-binimran/design-pack/skills/figma-handoff Skill

    3★ repo

    Pull a Figma design via the Figma connector and produce a developer handoff - tokens, specs, redlines, asset list, and component notes. Use when handing a design to engineering or auditing a Figma file.

  • Forge

    claudialnathan/agent-kitchen/skills/forge Skill

    no license3★ repo

    Designs coding-agent harness artifacts. One triage ladder picks the right surface (skill, hook, path-scoped rule, CLAUDE.md or AGENTS.md entry, workflow script, subagent, MCP); the stance holds the bar every artifact must clear (a gradeable objective, a gap between the model's default and what the owner wants, and a nameable attention shift; the gap is a new idea or source worth operationalizing, a miscommunication to prevent, or an observed failure); per-surface mechanics live in references. Use for creating, reviewing, or refactoring any of these, for distilling articles or docs into one, for skills that under- or over-trigger, hooks that don't fire or fire too often, bloated CLAUDE.mds, and for deciding which surface a behavior belongs on. Triggers on any work on a skill, hook, rule, CLAUDE.md, AGENTS.md, or workflow orchestration script, including debugging triggering or firing semantics and packaging a /command or plugin.

  • Harness audit

    claudialnathan/agent-kitchen/skills/harness-audit Skill

    no license3★ repo

    Audits the coding-agent harness a project's sessions run under, end-to-end: what loads every session, what it costs, whether the harness's own claims hold, and whether its parts work when assembled. Inventories the standing surfaces (CLAUDE.md chain, personal/project/plugin skills, MCP instruction blocks, agent rosters, hooks, settings at every scope), quantifies their per-session token cost, runs setup checks (self-consistency, duplication across scopes, enforcement parity, scope discipline, and a provisional routing-effectiveness check for doc-fleet repos), and after substantial harness changes can run a representative composition probe through the complete installed system. The audit itself is read-only: the deliverable is a ranked findings report leading with the highest-value edit, nothing is applied without the owner's explicit go-ahead, and edits stay inside the project directory. Machine scope (anything under ~/.claude, user or enterprise settings, global plugins) is never edited; those findings are reported with a project-scoped counterweight proposed where one exists. Use to review, slim, or sanity-check a project's harness after major harness changes or when sessions feel slow or behave incoherently from turn one; for a single artifact, the relevant forge owns the audit.

  • Ingest

    claudialnathan/agent-kitchen/skills/ingest Skill

    no license3★ repo

    Reads content the owner hands over and works out where it belongs in the harness: which existing primitive it improves, whether it warrants a new one, or that nothing does. Scoped to the primitives forge designs (skills, hooks, path-scoped rules, CLAUDE.md and AGENTS.md entries, workflow scripts, subagents, MCP), read for what they mean for how those behave rather than summarized. Sources a request is built on, or one author's body of work being distilled, are read whole in the main thread whatever the count; a large pile of side references may fan out to one subagent per source that returns quoted excerpts. Surveys whatever harness exists around it, a skills repo or a codebase's .claude or nothing, instead of assuming a shape, then surfaces the placement for the owner to confirm and packages it as a forge-ready brief. Use when handing over reading, links, or someone's public writing to turn into a harness improvement.

  • Pwnote offsec osai

    Pwnote/skills/skills/pwnote-offsec-osai Skill

    3★ repo

    Use whenever the user is working on Offsec's OSAI / AI Red Teaming certification track, or on AI/LLM/agentic security engagements generally — prompt injection findings, tool-use abuse, agent trajectory documentation, or writing up AI-specific security findings that don't map cleanly to traditional CVSS. Trigger on "OSAI", "AI Red Teaming", "LLM pentest", "prompt injection engagement", or "agentic AI exploitation", even without the word "skill".

  • Sync conversations

    yinjialu/context-harness/skills/sync-conversations Skill

    3★ repo

    Sync local Codex and Claude Code conversation records into a context-harness data home. Use when the user asks to back up Code Agent conversations, sync Codex sessions, sync Claude Code sessions, run full or incremental conversation backup, or inspect context-harness sync status.

  • Dispatching parallel agents

    EvezArt/evez-skills/skills/dispatching-parallel-agents Skill

    no license3★ repo

    Use when facing 2+ independent tasks that can be worked on without shared state or sequential dependencies

    The same file is in 15 other repositories

  • Skill creator

    EvezArt/evez-skills/skills/skill-creator Skill

    no license3★ repo

    Guide for creating effective skills. This skill should be used when users want to create a new skill (or update an existing skill) that extends Manus's capabilities with specialized knowledge, workflows, or tool integrations.

    The same file is in 11 other repositories

  • Skill creator

    heshengta0/super-agent-party/skills/skill-creator Skill

    3★ repo

    Guide for creating effective skills. This skill should be used when users want to create a new skill (or update an existing skill) that extends Claude's capabilities with specialized knowledge, workflows, or tool integrations.

    The same file is in 1 other repository

  • Hook writer

    m-binimran/creator-pack/skills/hook-writer Skill

    3★ repo

    Write scroll-stopping hooks - the first line / first 3 seconds - for Reels, carousels, and posts. Use when a piece needs a stronger opening or you want hook options to test.

  • Prompt master gpt5

    George930502/curated-codex-skills/skills/prompt-master-gpt5 Skill

    3★ repo

    Generate, audit, tighten, adapt, or decompile prompts for the GPT-5.6 model family. Use when the user wants a first-pass-ready GPT-5.6 prompt for an API, ChatGPT, Codex, coding agent, research, tool use, or structured output.