Examples catalog

Custom AI personas for specialized tasks. Place in .claude/agents/ or ~/.claude/agents/

• adr-writer.md Details →

  Architecture Decision Record generator: read-only, Opus

• analytics-with-eval/analytics-agent.md

  Analytics agent with LLM-as-a-judge evaluation loop for output quality

• analytics-with-eval/eval/report-template.md

  Monthly evaluation template for scoring analytics agent performance and accuracy

• architecture-reviewer.md Details →

  Architecture & design review: read-only, Opus, never modifies code

• code-reviewer.md Details →

  Thorough code review with best practices

• cyber-defense/anomaly-detector.md

  Detect security log anomalies using statistical baselines

• cyber-defense/log-ingestor.md Details →

  Ingest and normalize logs from multiple sources for threat analysis

• cyber-defense/risk-classifier.md

  Classify security events by severity using CVSS framework

• cyber-defense/threat-reporter.md

  Generate structured threat reports from classified security events

• devops-sre.md Details →

  Infrastructure troubleshooting with FIRE framework

• implementer.md Details →

  Mechanical execution agent: Haiku, bounded scope

• integration-reviewer.md Details →

  Runtime integration validator (read-only)

• loop-monitor.md

  Detect and interrupt runaway agent loops (read-only watchdog)

• output-evaluator.md

  LLM-as-a-Judge quality gate

• plan-challenger.md

  Adversarial plan review across 5 dimensions (read-only)

• planner.md Details →

  Strategic planning agent: read-only, Opus, use before implementation

• planning-coordinator.md

  Synthesis agent for dynamic research teams (read-only)

• refactoring-specialist.md

  Clean code refactoring expert

• security-auditor.md

  Security vulnerability detection (OWASP)

• security-patcher.md

  Apply security patches from audit findings, proposes for review

• test-writer.md

  TDD/BDD test generation specialist

Reusable knowledge modules. Place in .claude/skills/ or ~/.claude/skills/

• git-ai-archaeology/

  Analyze AI config evolution in a git repo: first commits, monthly distribution, maturity phases

• design-patterns/

  Detect and analyze GoF design patterns

• voice-refine/

  Writing voice refinement with before/after examples

• rtk-optimizer/

  RTK token optimization analysis

• audit-agents-skills/

  Quality audit for agents, skills, and commands

• skill-creator/

  Create new skills with proper structure and best practices

• landing-page-generator/

  Generate deploy-ready landing pages from any repository

• ccboard/

  TUI/Web dashboard for Claude Code session monitoring

• guide-recap/

  Transform CHANGELOG entries into social content

• release-notes-generator/

  Generate release notes in 3 formats from git commits

• pr-triage/

  4-phase PR backlog management with worktree setup

• issue-triage/

  3-phase issue backlog management

• cyber-defense-team/

  Multi-agent cyber defense team orchestration

• talk-pipeline/

  6-stage pipeline: raw material to slides via Kimi

• token-audit/

  Measure fixed-context token overhead, classify rules by usage frequency, audit hook cost

• eval-skills/

  Audit all skills for frontmatter completeness and effort-level inference

• eval-rules/ Details →

  Audit .claude/rules/: resolves glob patterns against real files, interactive usefulness review

• mcp-integration-reference/

  MCP integration reference with Sentry patterns and multi-tool query examples

• plan-pipeline/

  Complete planning pipeline: product direction to architecture to implementation plan, validation, and execution

• ast-grep-patterns.md

  AST-based code search patterns

• pdf-generator.md Details →

  Professional PDF generation (Quarto/Typst)

• security-checklist.md Details →

  OWASP Top 10 security checks

• smart-explore.md Details →

  Progressive code exploration via tree-sitter AST, 86-92% token reduction

• tdd-workflow.md

  Test-Driven Development process

• audit-codebase/

  Codebase health audit scoring 7 categories with progression plan

• autoresearch/

  Autonomous improvement loop: scan codebase metrics, scaffold experiments, run agent-driven iterations

• best-of-n/

  Generate bounded independent candidates, score them against a frozen rubric, and verify the selected result with a proof

• canary/

  Post-deploy monitoring: watch production after a deploy and alert on regressions

• catchup/

  Restore context after /clear by summarizing recent work and project state

• check-cache-bugs/

  Audit Claude Code setup for cache bugs CC#40524: sentinel, --resume, attribution header

• ci-all/ Details →

  Full CI pipeline: local tests, type check, push branch, return the pipeline URL

• ci-pipeline/

  Push current branch and return the pipeline tracking URL (GitLab or GitHub Actions)

• ci-status/

  Show current pipeline status for the active branch

• ci-tests/

  Run the test suite: auto-detects Python (pytest/uv), Node (vitest/pnpm), or Rust (cargo test)

• commit/ Details →

  Generate a conventional commit message for staged changes

• diagnose/

  Interactive troubleshooting assistant for Claude Code issues

• eval-agents/ Details →

  Audit agent definitions for description specificity, model tier, tools scoping, and dispatch ambiguity between agents

• eval-hooks/ Details →

  Audit hook definitions in settings.json for validity, performance safety, and exit-code strategy correctness

• explain/

  Explain code, concepts, or system behavior with adjustable depth levels

• generate-tests/

  Generate comprehensive tests for specified code

• git-worktree/

  Create isolated git worktrees for feature development without switching branches

• git-worktree-clean/

  Clean up stale git worktrees with merged branch detection and disk usage report

• git-worktree-remove/

  Safely remove a git worktree with branch cleanup and safety checks

• git-worktree-status/

  Check status of active worktrees and background verification tasks

• handoff-create/ Details →

  Generate a structured handoff capturing scope, relevant files, discoveries, work completed, and next steps

• handoff-resume/

  Load a handoff document and resume work from where a previous session left off

• handoff-update/

  Update an existing handoff: append-only Work Done history, replace Status and Next Steps

• investigate/ Details →

  Systematic root-cause debugging: find the cause before writing any fix

• land-and-deploy/ Details →

  Merge PR, wait for CI, verify deploy, run canary: the complete landing pipeline

• learn-alternatives/

  Compare different approaches to solve the same problem

• learn-quiz/

  Test understanding of recently written or accepted code

• learn-teach/

  Step-by-step explanation of a concept with progressive depth

• learning-path/

  Run the local Claude Code learning path, record evidence, and schedule evidence-based reviews.

• methodology-advisor/

  Analyze codebase and ask 3 targeted questions to recommend the right AI dev methodology stack

• optimize/

  Analyze and suggest performance improvements for code, queries, or systems

• pr/ Details →

  Analyze changes, detect scope issues, and create a well-structured PR

• qa/

  Systematic QA testing of a web app: diff-aware, tiered, with fix-and-verify loop

• recipe-template/

  Template for recipe-style commands with precondition validation pattern

• refactor/

  Analyze code for SOLID violations and suggest targeted improvements

• release-notes/

  Generate release notes in multiple formats from git commits

• review-plan/

  Structured plan review across 4 axes before writing any code

• review-pr/

  Perform a comprehensive code review of a pull request

• routines-discover/

  Analyze the project to surface high-value Routines use cases across schedule, API, and GitHub triggers

• sandbox-status/

  Display native sandbox status, configuration, and recent violations

• sandbox-unblock/

  Eight-check diagnostic protocol before reporting a sandbox blocker, plus a report template and escalation path

• scaffold/ Details →

  Interactive coach: asks 4-5 questions to determine the right template type then generates it

• security/

  Rapid security assessment focused on OWASP Top 10 vulnerabilities

• security-audit/

  Comprehensive security audit with scored posture assessment

• security-check/

  Quick configuration check against known threats database

• session-save/

  Save current session state (decisions, modified files, status, next steps) to a handoff file

• ship/ Details →

  Comprehensive pre-deployment verification to ensure release readiness

• sonarqube/

  Analyze SonarCloud quality issues for a specific PR

• update-threat-db/

  Research and update the AI agent security threat intelligence database

• validate-changes/

  Evaluate staged changes using LLM-as-a-Judge before committing

Custom slash commands (user-invocable skills). Place in .claude/skills/ with disable-model-invocation: true

• audit-agents-skills.md

  /audit-agents-skills - Quality audit for .claude/ config

• audit-codebase.md

  /audit-codebase - Codebase health audit scoring 7 categories

• autoresearch.md Details →

  /autoresearch - Autonomous improvement loop: scan metrics, scaffold, run agent iterations

• canary.md

  /canary - Post-deploy monitoring with regression alerts vs baseline

• catchup.md

  /catchup - Restore context after /clear

• check-cache-bugs.md Details →

  /check-cache-bugs - Audit for cache bugs CC#40524 (sentinel, --resume, attribution header)

• ci/all.md Details →

  /ci:all - Full CI: local tests + type check + push + pipeline URL

• ci/pipeline.md

  /ci:pipeline - Push branch and return pipeline tracking URL (GitLab/GitHub Actions)

• ci/status.md

  /ci:status - Current pipeline status for active branch

• ci/tests.md

  /ci:tests - Auto-detect and run test suite (pytest/vitest/cargo test)

• commit.md Details →

  /commit - Conventional commit messages

• diagnose.md

  /diagnose - Interactive troubleshooting

• explain.md

  /explain - Code explanations (3 depth levels)

• generate-tests.md

  /generate-tests - Test generation

• git-worktree-clean.md

  /git-worktree-clean - Batch cleanup of stale worktrees

• git-worktree-remove.md

  /git-worktree-remove - Safe worktree removal with merge checks

• git-worktree-status.md

  /git-worktree-status - Check background verification tasks

• git-worktree.md Details →

  /git-worktree - Isolated worktree setup

• handoff/create-handoff.md

  /create-handoff - Generate structured handoff: scope, files, discoveries, next steps

• handoff/resume-handoff.md

  /resume-handoff - Load handoff doc and resume work from previous session

• handoff/update-handoff.md

  /update-handoff - Update handoff with append-only Work Done history

• investigate.md Details →

  /investigate - Systematic root-cause debugging before writing any fix

• land-and-deploy.md Details →

  /land-and-deploy - Merge PR → CI → verify deploy → canary pipeline

• learn/alternatives.md

  /learn:alternatives - Compare different approaches

• learn/quiz.md

  /learn:quiz - Self-testing for learning concepts

• learn/teach.md

  /learn:teach - Step-by-step concept explanations

• methodology-advisor.md

  /methodology-advisor - Analyze codebase + 3 questions → recommended AI dev methodology stack

• optimize.md

  /optimize - Performance analysis and roadmap

• plan-ceo-review.md

  /plan-ceo-review - Strategic product gate: find the 10-star product before writing code (gstack)

• plan-eng-review.md

  /plan-eng-review - Engineering gate: lock architecture, diagrams, edge cases before coding (gstack)

• plan-execute.md Details →

  /plan-execute - Execute plan with TDD, parallel agents, PR creation

• plan-start.md Details →

  /plan-start - 5-phase planning with dynamic research team

• plan-validate.md Details →

  /plan-validate - 2-layer validation with specialist agents

• pr.md Details →

  /pr - Create well-structured PRs

• qa.md

  /qa - Diff-aware browser QA: 3 tiers with fix-and-verify loop

• recipe-template.md

  Template for recipe-style commands with precondition validation pattern

• refactor.md

  /refactor - SOLID-based code improvements

• release-notes.md Details →

  /release-notes - Generate in 3 formats

• resources/threat-db.yaml

  Threat intelligence database for security commands

• review-plan.md

  /review-plan - Structured plan review across 4 axes

• review-pr.md

  /review-pr - PR review workflow

• routines-discover.md

  /routines-discover - Scan project and surface high-value Routines use cases (schedule/API/GitHub triggers)

• sandbox-status.md

  /sandbox-status - Sandbox isolation status check

• scaffold.md Details →

  /scaffold - Interactive coach: generates agent/command/skill/hook/rule template

• security-audit.md

  /security-audit - Full 6-phase audit with score /100

• security-check.md Details →

  /security-check - Config scan vs known threats (~30s)

• security.md

  /security - Quick OWASP security audit

• session-save.md

  /session-save - Save session state (decisions, files, next steps) to handoff file

• ship.md

  /ship - Pre-deploy checklist

• sonarqube.md

  /sonarqube - Analyze SonarCloud issues

• update-threat-db.md

  /update-threat-db - Research & update threat intelligence

• validate-changes.md

  /validate-changes - Pre-commit validation

Event-driven automation scripts for macOS/Linux. Place in .claude/hooks/

• auto-checkpoint.sh

  Auto-checkpoint work at intervals

• auto-format.sh

  Auto-format after edits (Prettier, Black, go fmt)

• auto-rename-session.sh Details →

  AI-powered session title generation (Haiku)

• claudemd-scanner.sh

  Scan CLAUDE.md for injection attacks

• dangerous-actions-blocker.sh Details →

  Block dangerous commands/edits

• file-guard.sh

  Protect sensitive files from modification

• governance-enforcement-hook.sh

  Enforce project config, MCP registry, and data classification policies

• identity-reinjection.sh

  Re-inject agent identity after context compaction

• learning-capture.sh

  Prompt for daily learning capture

• mcp-config-integrity.sh

  Verify MCP config hash (CVE protection)

• notification.sh

  macOS contextual sound alerts

• output-secrets-scanner.sh

  Detect secrets in tool outputs

• output-validator.sh

  Heuristic output validation

• permission-request.sh

  Explicit permission flow for risky ops

• pre-commit-evaluator.sh

  LLM-as-a-Judge pre-commit validation

• pre-commit-secrets.sh

  Block secrets from entering commits

• privacy-warning.sh

  Warn on privacy-sensitive ops

• prompt-injection-detector.sh Details →

  Detect injection attempts

• repo-integrity-scanner.sh

  Scan README/package.json for injection

• rtk-auto-wrapper.sh

  Auto-wrap commands with RTK for token savings

• rtk-baseline.sh

  Save RTK baseline for session savings tracking

• sandbox-validation.sh

  Validate sandbox isolation

• security-check.sh

  Block secrets in commands

• security-gate.sh

  Detect vulnerable code patterns before writing

• session-logger.sh

  Log operations for monitoring

• session-summary-config.sh

  Configure session-summary sections and display

• session-summary.sh Details →

  Full session analytics (15 configurable sections)

• setup-init.sh

  Initialize session environment

• smart-suggest.sh Details →

  Suggest relevant commands/agents based on prompt intent

• subagent-stop.sh

  Clean up sub-agent resources

• test-hooks.sh

  set -euo pipefail

• test-on-change.sh

  Run tests on file changes

• tts-selective.sh

  Text-to-speech for selected outputs

• typecheck-on-save.sh

  Run TypeScript checks on save

• unicode-injection-scanner.sh

  Detect zero-width, RTL, ANSI escape

• velocity-governor.sh

  Rate-limit tool calls to avoid API throttling

• verification-gate.sh

  set -euo pipefail

Event-driven automation scripts for Windows

• auto-format.ps1

  Auto-format after edits

• security-check.ps1

  Block secrets in commands

Configuration file templates. Place in .claude/ or ~/.claude/

• CONTRIBUTING-ai-disclosure.md

  AI disclosure template for CONTRIBUTING.md

• PULL_REQUEST_TEMPLATE-ai.md

  PR template with AI attribution

• mcp.json

  MCP servers configuration

• sandbox-native.json Details →

  Native Claude Code sandbox configuration

• settings-personalization.json

  UI personalization: spinner verbs, custom tips

• settings.json

  Hooks and preferences config

CLAUDE.md memory file templates for persistent context

• icm-session-starter.md

  ICM session bootstrap: recall past decisions and store new ones

Ready-to-use CLAUDE.md configuration profiles. Add to ~/.claude/CLAUDE.md or project CLAUDE.md

• TESTING.md

  Copy this file into a work item before testing a change, an agent workflow, or a Best-of-N selection. Fill in fields wit

• agent-progress.md

  <!--

• design-reference-file.md

  Brand-book and UI kit context for consistent UI generation

• devops-sre.md

  DevOps/SRE project configuration

• learning-mode.md

  Learning-focused development configuration

• product-designer.md

  Product designer workflow configuration

• rtk-optimized.md

  RTK token-optimized configuration

• session-naming.md Details →

  Auto-rename sessions with descriptive titles for parallel work

• tts-enabled.md

  Text-to-speech enabled configuration

Behavioral rules auto-loaded by Claude for common review patterns. Place in .claude/rules/

• architecture-review.md

  Rules for architecture review sessions

• code-quality-review.md

  Rules for code quality review

• first-principles.md Details →

  First-principles reasoning rules

• performance-review.md

  Rules for performance review sessions

• test-review.md

  Rules for test review sessions

Utility scripts for setup, diagnostics, and monitoring

• ai-usage-charter-template.md

  Team AI usage charter template: principles, scope, review cadence

• audit-scan.sh

  Fast setup audit scanner

• bridge-plan-schema.json

  JSON Schema for bridge plan v1 format

• bridge.py

  Bridge: Claude Code plan → LM Studio execution

• cc-sessions.py Details →

  Advanced session search with incremental indexing

• check-claude.ps1

  Health check (Windows)

• check-claude.sh

  Health check (macOS/Linux)

• clean-reinstall-claude.ps1

  Clean reinstall (Windows)

• clean-reinstall-claude.sh

  Clean reinstall (macOS/Linux)

• fresh-context-loop.sh

  Auto-restart sessions at context limits

• mcp-registry-template.yaml

  MCP server registry template for team governance

• migrate-arguments-syntax.ps1

  Migrate v1 → v2 slash command argument syntax (PowerShell)

• migrate-arguments-syntax.sh

  Migrate v1 → v2 slash command argument syntax (bash)

• og-image-astro.ts

  Generate OG images for Astro static sites

• pptx-to-pdf.sh

  Batch convert PPTX to PDF via Keynote (macOS/AppleScript)

• rtk-benchmark.sh

  Benchmark RTK token savings vs raw commands

• session-search.sh

  Search & resume sessions

• session-stats.sh

  Analyze session logs & costs

• smart-suggest-roi.py Details →

  Measure SmartSuggest hook ROI: track suggestion follow-through rates

• sonnetplan.sh

  Run Claude with Sonnet replacing Opus (cost optimization)

• statusline.py

  Terminal statusline for Claude Code session monitoring

• supply-chain-triage.py

  Triage a workstation after an npm supply chain incident

• sync-claude-config.sh

  Sync Claude config files across machines

• test-prompt-caching.ts

  Test prompt caching implementation and measure savings

Templates for scaling Claude Code across teams. Place in .claude/

• claude-skeleton.md

  Minimal CLAUDE.md skeleton for new team members

• profile-template.yaml

  Profile assembly template for multi-tool teams

• sync-script.ts

  Sync Claude config across team machines

Session and workflow templates for context continuity

• ai-assisted-contribution.md

  Record reproduction, verification evidence, author explanation and maintainer follow-up

• feature-list.json

  Claude Code harness scope tracker: active features with WIP=1 constraint and evidence fields

• session-handoff-lorenz.md

  Session handoff template for context continuity across sessions

CI/CD workflows for GitHub Actions automation

• .coderabbit.yaml

  CodeRabbit AI code review configuration, copy to repo root

• .greptile/rules.md

  Greptile cross-file invariants rulebook template

• claude-code-review-batched.yml

  Domain-split matrix code review for PRs over a file-count threshold

• claude-code-review.yml Details →

  Prompt-based review with anti-hallucination protocol

• claude-issue-triage.yml

  Auto-triage issues with labels

• claude-pr-auto-review.yml

  Auto code review on PRs

• claude-security-review.yml

  Security-focused PR scan

• prompts/code-review.md

  Code review prompt template for GitHub Actions workflows

Advanced development workflow guides

• bounded-loop-contract.md

  Reference contract for a bounded agent loop with a separate verifier interface, evidence, and escalation

• database-branch-setup.md

  Isolated feature dev with DB branches (Neon/PlanetScale)

• feature-list.json

  Harness Engineering scope tracker: one active feature at a time with evidence-based status

• lean-harness-experiment.md

  Test one repository harness countermeasure against a known failure and its operating cost

• memory-stack-integration.md

  Multi-day workflow with memory tools (claude-mem + Serena + grepai)

• remotion-quickstart.md

  Video generation workflow with Remotion

• review-admission.md

  Define shared review capacity, pause and resume rules, and a tabletop exercise

Community plugins extending Claude Code capabilities

• claude-mem.md

  Persistent memory management plugin

• code-quality/plugin.json

  {

• devops-pipeline/plugin.json

  {

• pr-workflow/plugin.json

  {

• release-automation/plugin.json

  {

• se-cove.md

  Chain-of-Verification for independent code review (Meta AI, ACL 2024)

• security-suite/plugin.json

  {

Community integrations and MCP server extensions for Claude Code

• agent-vibes/README.md

  Agent Vibes: text-to-speech MCP server overview and features

• agent-vibes/installation.md

  Install Agent Vibes TTS on macOS step-by-step

• agent-vibes/troubleshooting.md

  Troubleshoot Agent Vibes TTS installation issues

• agent-vibes/voice-catalog.md

  Agent Vibes voice catalog reference

MCP server configuration files

• figma.json

  Figma MCP server configuration

Behavioral modes for Claude (SuperClaude framework). Place in ~/.claude/

• MODE_Learning.md Details →

  Just-in-time explanations mode

Templates for structuring Claude's context: code maps, budgets, eval questions, CI drift checks

• assembler.ts

  Assemble CLAUDE.md from profile YAML + shared module markdown files

• canary-check.sh

  Verify context integrity between sessions

• ci-drift-check.yml

  CI workflow to detect context drift on each push

• code-map-template.yaml

  Structural metadata file: codebase shape in ~1K tokens

• context-bench.sh

  Benchmark structural metadata impact on context quality

• context-budget-calculator.sh

  Calculate token budget across all context sources

• eval-questions.yaml

  Evaluation questions for context quality scoring

• profile-template.yaml

  Context profile template: env, stack, constraints

• rules/knowledge-feeding.md

  Knowledge feeding protocol: inject domain context systematically

• rules/update-loop-retro.md

  Session retrospective template for capturing and persisting learnings

• skeleton-template.md Details →

  Minimal context skeleton: project structure, key files, current focus

Custom style templates for consistent AI output tone and formatting

• custom-style-template.md

  Define a custom writing style for Claude: tone, vocabulary, formatting rules

Curated technical terms with examples, sources, and usage limits

• anchor-catalog.md Details →

  Curated technical vocabulary with examples, sources, and usage limits