Context & AI engineering
25,610 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
Swift ui design iteration mcp loop
esaldgut/ai-native-engineering-workspace/global-skills/apple/swift-ui-design-iteration-mcp-loop Skill
0★ repoA methodology for iterating SwiftUI UI with a visual, MCP-driven feedback loop — write a #Preview, render it headlessly, screenshot the running view, evaluate the bitmap against Apple's Human Interface Guidelines, then refine the code and repeat, keeping code + screenshot + evaluation on one commit so the loop doesn't drift. Encodes the HIG rules to check each pass (contrast, Dynamic Type, Reduce Transparency/Motion, touch targets, "glass on controls, not content"), a SwiftUI performance checklist (state granularity, no GeometryReader in lazy stacks, .drawingGroup for static layers), and which MCP tools do which step. Use when designing or polishing a view and you want grounded visual evaluation instead of guessing from code.
Claude mcp lesson capture pipeline
esaldgut/ai-native-engineering-workspace/global-skills/claude-code-workflow/claude-mcp-lesson-capture-pipeline Skill
0★ repoCapture lessons from a merged PR and propagate them in a cascade — auto-memory, then agent skills, then human docs — so hard-won knowledge (compiler quirks, security findings, framework bugs) is never re-learned. Built on Claude Code's auto memory, a Stop hook (exit 0 + decision:block to continue), and a side-effectful skill marked disable-model-invocation so it runs only on /capture-lessons. Use after a merge to a long-lived branch, when a PR title matches chore(lessons): capture PR#N, or when asked to record lessons from a diff. This is the generic cross-platform pattern; platform-specific variants (e.g. an Android cascade) exist separately.
esaldgut/ai-native-engineering-workspace/global-skills/claude-code-workflow/cli-build-sync-pipeline Skill
0★ repoRun a platform-agnostic project-consistency pass before a commit/PR — regenerate the project file, format, lint, build, then take an IDE second opinion via the editor's MCP server to catch diagnostics the CLI build truncates or never emits. Works for Xcode, Gradle, Node, Go, etc. Uses Claude Code's PostToolUse hook (format-on-write) and the IDE-Navigator check (e.g. Xcode's XcodeListNavigatorIssues) as a complement to the build log. Use before declaring a change ready, on /sync, or when wiring format/lint/build automation. Generic cross-platform version; an Android-specific build-sync variant exists separately.
esaldgut/ai-native-engineering-workspace/global-skills/claude-code-workflow/mcp-orchestration-pattern Skill
0★ repoAssemble multiple specialized MCP servers so each owns one primary domain, with a task→MCP routing table in CLAUDE.md, a Phase 0 sweep before architecture, and Tool Search left on by default (deferral) instead of alwaysLoad-ing everything. Avoids the tool-overflow problem where Claude has 150 tools and picks badly. Uses Claude Code's MCP scope precedence (local > project > user), .mcp.json, alwaysLoad, and the MCP spec's server "instructions". Use when wiring more than one MCP server, when Claude isn't using an installed server, or when designing a cross-domain task (build + docs + design + backend) that spans several servers.
andy-builds-ai/claude-code-skills/mcp-bau Skill
0★ repoWhen building or extending an MCP server. Build one tool at a time, and prove every tool on the test bench (plain call, then Inspector) before it enters a client config.
detro/quirks/skills/ground-rules Skill
0★ repoHelps setup, view, and edit a global context or "ground rules" file for your AI agents (such as Crush, Claude Code, or Cursor), including configuring custom global context file paths and auto-detecting the executing agent.
K-9Nine/skipthespec/eval-loops Skill
0★ repoDesign, build, run, and maintain automated test and evaluation loops so Claude can verify its own work without deploying to users. Use when the user says 'eval loop', 'set up evals', 'test harness', 'automated testing', 'how do I know it works', 'regression suite', 'grade the agent', or when a build needs a verification net. Implements Anthropic's evals method: 20-50 tasks from real failures, deterministic graders + LLM-as-judge, grade outcomes not paths, isolated trials, pass@k vs pass^k, read the transcripts. The safety net under autonomous-build and the quality gate before dogfood-loop promotion. Load autonomous-dev first.
whyy9527/codexception/examples/codex-hook-json-sessionstart Skill
0★ repoFix Codex SessionStart hooks that print plain text, invalid JSON, or reminders that do not enter model-visible context. Use when editing ~/.codex/hooks.json, ~/.codex/hooks/*.sh, or a skill-bundled hook script for Codex.
lensetek/Digital-Marketing-Agent_Skills/capabilities/knowledge-base-rag Skill
0★ repoDigital-Marketing-Agent_Skills
lensetek/Digital-Marketing-Agent_Skills/capabilities/project-memory-learning Skill
0★ repoDigital-Marketing-Agent_Skills
lensetek/Digital-Marketing-Agent_Skills/skills/agent-update-manager Skill
0★ repoChecks repository changes, reviews changelogs, assesses compatibility, and proposes approval-gated updates.
Competitive intelligence agent
lensetek/Digital-Marketing-Agent_Skills/skills/competitive-intelligence-agent Skill
0★ repoMonitors competitor offers, pricing, campaigns, SEO activity, and industry benchmarks.
lensetek/Digital-Marketing-Agent_Skills/skills/crm-customer-retention-agent Skill
0★ repoDevelops segmentation, loyalty, repeat-purchase, churn-prevention, and win-back strategies.
lensetek/Digital-Marketing-Agent_Skills/skills/growth-creative-agent Skill
0★ repoGenerates unconventional campaign ideas, viral concepts, meme marketing, hooks, and creative experiments.
lensetek/Digital-Marketing-Agent_Skills/skills/marketing-orchestrator Skill
0★ repoMain entry point that receives briefs, plans workflows, routes work to specialists, manages approvals, and compiles the final report.
lensetek/Digital-Marketing-Agent_Skills/skills/social-media-engagement-agent Skill
0★ repoPlans captions, social calendars, community engagement, audience interaction, and organic growth.
krkrkrr/skills/skills/meta/retrospective-codify Skill
no license0★ repoPair "what failed first" with "what finally worked" and codify the should-have-known-it insight as an ast-grep rule, a skill, or a CLAUDE.md rule. Meta-skill: invoke ONLY when the user explicitly says "codify today's lessons," "make it a skill," "drop it into lint," or asks to extract a reusable rule from a trial-and-error fix. Do NOT auto-invoke at every task completion. Spares future-you (or another agent) the same trap.
krkrkrr/skills/skills/meta/skill-finder Skill
no license0★ repoUse ONLY when the user explicitly asks to discover or evaluate a skill from OUTSIDE the curated catalog — e.g., "find a Stripe skill", "is there a skill for X?", "evaluate this candidate before adopting". Meta-skill, complementary to skill-selector: do NOT auto-invoke on routine apm-management, project-init, or catalog-resident requests. Cross-source survey across vetted registries (Anthropic official, claude-skill-registry, VoltAgent/awesome-agent-skills, ComposioHQ, Superpowers, GitHub topic) with a mandatory waxa-eval adoption gate.
vincendiary/agents/skills/handoff Skill
no license0★ repoCompact the current conversation into a handoff document for another agent to pick up.
The same file is in 13 other repositories
vincendiary/agents/skills/setup-matt-pocock-skills Skill
no license0★ repoSets up an `## Agent skills` block in AGENTS.md/CLAUDE.md and `docs/agents/` so the engineering skills know this repo's issue tracker (GitHub or local markdown), triage label vocabulary, and domain doc layout. Run before first use of `to-issues`, `to-prd`, `triage`, `diagnose`, `tdd`, `improve-codebase-architecture`, or `zoom-out` — or if those skills appear to be missing context about the issue tracker, triage labels, or domain docs.
The same file is in 15 other repositories
vincendiary/agents/skills/write-a-skill Skill
no license0★ repoCreate new agent skills with proper structure, progressive disclosure, and bundled resources. Use when user wants to create, write, or build a new skill.
The same file is in 20 other repositories
agent-packs/registry/skills/ai-feature-development Skill
no license0★ repoBuild AI-powered features with disciplined prompt engineering, evaluation, and observability. Use when integrating LLMs, building agents, or shipping AI capabilities.
agent-packs/registry/skills/context-engineering Skill
no license0★ repoLoad and preserve the right project context at the right time. Use when starting sessions, switching tasks, or improving agent output quality.
agent-packs/registry/skills/using-agent-skills Skill
no license0★ repoSelect and sequence applicable Agent Skills for a coding session. Use when deciding which workflow skills should guide an agent's work.
mfraunhofer/claude-skills/usage-dashboard Skill
0★ repoLocal token-usage dashboard for Claude Code. Parses all session transcripts (~/.claude/projects/**/*.jsonl), aggregates usage per day / model / project / session / time-of-day / main-chat-vs-subagent, and renders a self-contained HTML dashboard with a USD cost equivalent (API list prices). Use when you want to see how much you've spent, what a day/project/model cost, or your token usage over time. NOT for the rate-limit status line (the harness renders that itself).
bobmmao/agent-workflow-skills/skills/governance/memory-dream Skill
0★ repoMemory consolidation protocol. Reviews all stored memories, merges duplicates, removes noise and credentials, rewrites unclear entries, and enforces TTL expiration. Use when the user asks to clean up, consolidate, or review their memories. Also triggers automatically after sufficient activity (configurable).
bobmmao/agent-workflow-skills/skills/governance/memory-triage Skill
0★ repoALWAYS USE THIS SKILL. Persistent long-term memory protocol powered by mem0. After EVERY response, evaluate the conversation for durable facts worth storing via memory_store. Handles identity, preferences, decisions, configurations, rules, projects, and relationships. Use memory_store (NOT workspace files) for all user facts. Read this skill at the start of every session.
bobmmao/agent-workflow-skills/skills/growth/lead-handoff Skill
0★ repoValidate qualified distributor leads and write them into the shared approved_leads.jsonl handoff file for the sales agent
bobmmao/agent-workflow-skills/skills/tools/agent-browser Skill
0★ repoA fast Rust-based headless browser automation CLI with Node.js fallback that enables AI agents to navigate, click, type, and snapshot pages via structured commands.
The same file is in 2 other repositories
frankxai/mind-palace-agent-skills/skills/memory-palace-architect Skill
0★ repoDesign a memory palace — choose a real or imagined locus, lay a fixed non-crossing route, and space the stations. Use when the user says "design a memory palace", "pick a place to memorize in", "lay out my palace", "map a route for N items", or needs a structure before encoding. Writes the rooms/stations skeleton of palace.json for loci-encoder to fill.
frankxai/mind-palace-agent-skills/skills/number-memory Skill
0★ repoMemorize digits, dates, phone numbers, and playing cards by converting numbers into images. Use when the user says "memorize this number", "remember these dates", "how do I memorize a deck of cards", "PAO system", "Major System", or "Dominic System". Supplies the encoding for numeric targets, then hands the resulting images to loci-encoder for placement.
payne0420/claude-ultracode-agents/llm-endpoint Skill
0★ repoCall a configured OpenAI-compatible (`/v1/chat/completions`) or Anthropic-compatible (`/v1/messages`) HTTP endpoint as a lightweight model backend — a raw completion call, not an autonomous agent. Use as an alternative to the codex/cursor/opencode CLI harnesses when you want to hit an arbitrary model behind a gateway/router/proxy (e.g. for a second opinion, a quick review of pasted/piped context, or text generation), or to plug such an endpoint into an ultracode workflow. The base URL and API key are read from local config and are never hardcoded. Works with any `/v1` chat or `/v1/messages` endpoint. Triggers: "use my llm endpoint", "call my openai-compatible endpoint", "hit /v1/chat/completions", "use the /v1/messages endpoint", "use my model gateway / router / proxy", "second opinion from my endpoint", "run this through my API endpoint".
payne0420/claude-ultracode-agents/ultracode-external-agents Skill
0★ repoAuthor ultracode workflows whose agent() steps are executed by an EXTERNAL coding CLI — OpenAI Codex, Cursor, or opencode — instead of (or alongside) regular Claude subagents. Use this to run an individual agent() step on a different model/provider; that one bridged step drops into ANY workflow shape — a single second-opinion step in a pipeline, one off-budget implementation step, a design probe, or (when you want disagreement) a fan-out panel. The shape is the Workflow tool's job; this skill only changes what one agent() runs on. Explains the bridge (a thin Claude adapter shells out to the CLI), gives copy-paste delegation helpers, per-backend invocation recipes, structured-output mapping, worktree isolation for parallel writes, and a full runnable example workflow. Applies only inside an ultracode/Workflow authoring turn — the Workflow tool must be active; if it is not, this skill has nothing to do and should not be selected. Triggers: "use codex/cursor/opencode in a workflow", "ultracode with codex", "run my workflow on GPT/Composer instead of Claude", "multi-model workflow", "delegate workflow agents to an external CLI", "multi-model review panel", "external agents in a workflow", "fan out to different models".
stevancris/sre-ai-agent/skills/oncall-handoff Skill
0★ repoGenerate a structured on-call handoff document when shifting between on-call engineers. Use at the end of an on-call shift, when handing over an active incident to another engineer, or when an engineer needs to step away mid-incident. Ensures no context is lost between shifts and the incoming SRE can continue without starting from scratch. Trigger keywords: handoff, hand off, hand over, shift change, end of shift, my shift is ending, taking over, I'm stepping off, passing the pager, shift handover, oncall handover, who's next, briefing the next person, context transfer, shift summary.
TheophilusChinomona/idev/skills/skill-benchmark Skill
0★ repoEvaluates and benchmarks Agent Skill quality via static analysis and evaluation methodology. Use when assessing or reviewing skill quality, benchmarking skills, measuring skill activation accuracy, optimizing skill descriptions, or running /idev:benchmark-skills.
mike623/agent-skills/skills/life-agent-modes Skill
no license0★ repoUse when an agent should act as a general life assistant with persistent task modes, choosing the lightest suitable mode instead of over-applying one specialist mindset.
scchearn/loam/skills/loam-memory/loam-adding-to-memory Skill
0★ repoRead a local source file or synthesize conversation context, then integrate admitted content directly into topic, entity, concept, and analysis pages in existing memory (the wiki substrate). Use this when the user wants to add a source to the wiki, add a document, ingest a local note, transcript, article, report, or PDF, or explicitly preserve the current conversation as a topic note. For session-learning routing across wiki, guidance, checkpoint, task annotation, or discard, use /loam::learning-from-session. Must not ingest a goal wholesale; admit only independently reusable findings.
scchearn/loam/skills/loam-memory/loam-amending-memory Skill
0★ repoCorrect or update existing wiki content when newer evidence shows the wiki is wrong, stale, incomplete, or contradicted. Use this when the agent discovers the wiki says X but we now know Y, when code or real-world changes invalidate a wiki claim, or when the user asks to fix or amend the wiki. Not for adding new sources, routine learnings capture, structural normalization, or health checks; use /loam::adding-to-memory, /loam::learning-from-session, /loam::normalizing-memory, or /loam::linting-memory.
scchearn/loam/skills/loam-memory/loam-auditing-guidance Skill
0★ repoAudit, prune, and improve agent guidance markdown files in repositories. Use when the user asks to check, audit, update, improve, or fix AGENTS.md, CLAUDE.md, or related guidance files. Adds missing commands and gotchas, removes stale entries, deduplicates, and keeps the file small and relevant. Scan for guidance files, evaluate quality against templates, output a quality report, then make targeted updates.
scchearn/loam/skills/loam-memory/loam-linting-memory Skill
0★ repoRun a health check on existing memory (the wiki substrate) and goal artifacts. Use this when the user wants to lint the wiki, health-check the knowledge base, find orphan pages, spot broken or missing cross-links, clean up stale claims and unresolved wikilinks with safe local fixes, consolidate a legacy root `overview.md` into `index.md`, or health-check goal artifacts. Not for adding new material; use /loam::adding-to-memory or /loam::learning-from-session for that.
scchearn/loam/skills/loam-memory/loam-normalizing-memory Skill
0★ repoInspect an existing memory corpus (wiki substrate) and align it to this repo's Obsidian-friendly note-graph conventions. Use this when the user wants to import, normalize, retrofit, or clean up existing memory, notes folder, vault, docs tree, or mixed markdown knowledge base. In monorepos, also use it to align relevant AGENTS.md and CLAUDE.md files. Excludes goals/ from normalization. Not for routine wiki maintenance; use /loam::linting-memory for that.
scchearn/loam/skills/loam-memory/loam-querying-memory Skill
0★ repoAnswer questions against existing memory (the wiki substrate). Use this whenever the user is asking what is happening in the project, directory, codebase, architecture, workflow, decisions, or current state and the wiki likely contains the answer, even if they do not explicitly mention the wiki. Also use it for summaries, comparisons, and reusable analyses grounded in current wiki pages. Routes authoritative goal-state questions to /loam::setting-goals. Not for surfacing unresolved gaps; use /loam::reviewing-memory for that.
scchearn/loam/skills/loam-memory/loam-reviewing-memory Skill
0★ repoSurface and classify all open questions, unresolved contradictions, stale claims, and knowledge gaps from existing memory. Use this when the user wants to see what's still open, what needs attention, what's unresolved, or what still needs research or ingest. Does not classify incomplete goals as memory gaps. Not for answering questions, fixing issues, or adding new material; use /loam::querying-memory, /loam::linting-memory, /loam::amending-memory, /loam::adding-to-memory, or /loam::learning-from-session.
Neznakometz/StackForge/core/skills/memory Skill
0★ repoThe agent's memory between sessions — context management (L1) and checkpoints (L2). Apply when the context fills up, when moving to a new chat, and when closing a session.
Neznakometz/StackForge/core/skills/prompt-audit Skill
0★ repoAudit of the system prompt/CLAUDE.md against 12 criteria with a concrete improvement plan, plus a supply-chain safety screen for skills/agents you author or install. Apply to check the quality of your own sets and the project's CLAUDE.md, and before installing a third-party skill.
Neznakometz/StackForge/storage/skills/durable-memory Skill
0★ repoL3 durable memory for the agent across sessions — SQLite + agent-memory.md, consolidation and cleanup. Apply when you need persistent project memory (patterns, paths, architectural decisions) beyond the core checkpoints (L2).
Neznakometz/StackForge/storage/skills/rag Skill
0★ repoRetrieval-Augmented Generation — chunking, embeddings, vector storage, search and reranking. Apply when building search over knowledge/documents, Q&A over a corpus, semantic search.
khymerao/cursor-seo/skills/seo-dataforseo Skill
0★ repoLive SEO data via DataForSEO MCP server. SERP analysis (Google, Bing, Yahoo, YouTube, Google Images), keyword research (volume, difficulty, intent, trends), backlink profiles, on-page analysis (Lighthouse, content parsing), competitor analysis, content analysis, business listings, AI visibility (ChatGPT scraper, LLM mention tracking), and domain analytics. Requires DataForSEO extension installed. Use when user says "dataforseo", "live SERP", "keyword volume", "backlink data", "competitor data", "AI visibility check", "LLM mentions", "image SERP", "google images", "image rankings", or "real search data".
ucsandman/claude-code-capability-primer/skills/claude-code-capabilities Skill
0★ repoUse when unsure which Claude Code capability fits a task, when about to single-thread large/cross-cutting/repetitive work, or when you or the user ask what Claude Code can do — loads the full capability map (signal → capability) and points to the deep-dive skills.
ucsandman/claude-code-capability-primer/skills/claude-code-checkpointing Skill
0★ repoUse when you need to undo code edits, explore alternatives without losing a starting point, recover from a bad edit path, or restore previous conversation state — press Esc Esc or run /rewind to open checkpoints menu.
ucsandman/claude-code-capability-primer/skills/claude-code-dynamic-workflows Skill
0★ repoUse when a task is too big for one context — broad audits, multi-file migrations, multi-source research, or parallel multi-dimension review — and you want to fan out and orchestrate many subagents deterministically with the Workflow tool (requires ultracode / explicit opt-in).
ucsandman/claude-code-capability-primer/skills/claude-code-effort-models Skill
0★ repoUse when deciding how to work on a task - choosing effort level, model, or fast mode to balance reasoning depth, speed, and token cost.
ucsandman/claude-code-capability-primer/skills/claude-code-github-actions Skill
0★ repoUse when setting up or debugging Claude Code GitHub Actions automation, or deciding whether to route CI/CD tasks to GitHub Actions vs local session.
ucsandman/claude-code-capability-primer/skills/claude-code-hooks Skill
0★ repoUse when configuring automated lifecycle commands (hooks) in Claude Code settings.json — event types, matcher syntax, exit codes, JSON output, path placeholders.
ucsandman/claude-code-capability-primer/skills/claude-code-mcp Skill
0★ repoUse when configuring MCP servers in Claude Code — connecting external tools (GitHub, Stripe, Sentry, databases), choosing transports (HTTP, stdio, WebSocket), managing scopes, handling authentication, scaling with tool search, and decidi...
ucsandman/claude-code-capability-primer/skills/claude-code-plugins Skill
0★ repoUse when building, debugging, or installing Claude Code plugins—understanding plugin structure, CLI commands, marketplace configuration, or skill/agent packaging.
ucsandman/claude-code-capability-primer/skills/claude-code-skills Skill
0★ repoUse when you need to understand, create, trigger, or debug Claude Code skills—reusable workflows defined in SKILL.md files with YAML frontmatter.
ucsandman/claude-code-capability-primer/skills/claude-code-subagents Skill
0★ repoUse when you have research, exploration, a large review, or 2+ independent tasks to run — spawn subagents (the Agent/Task tool), in parallel when possible, to do focused work and keep the main context clean.
zajalist/zajal-skills/skills/lens-studio-mcp-editing Skill
0★ repoUse when editing a Lens Studio scene through the MCP server — setting properties, wiring component/asset references, attaching scripts, or navigating the scene graph. Covers the component-by-component rule, the reference "{}" echo, and script attachment.
The same file is in 1 other repository
goharabbas321/zeoel/skills/agent-architecture-audit Skill
no license0★ repoFull-stack diagnostic for agent and LLM applications. Audits the 12-layer agent stack for wrapper regression, memory pollution, tool discipline failures, hidden repair loops, and rendering corruption. Produces severity-ranked findings with code-first fixes. Essential for developers building agent applications, autonomous loops, or any LLM-powered feature.
The same file is in 3 other repositories