The library

write-a-skill · Teams and individuals building reusable workflows in Claude Code who need to document skills in a format that enables agent discovery and execution.

eng-runbook · Quick reference during on-call rotations when incident response must be immediate and procedures must be precise and copy-paste-ready.

standup · Engineers and PMs who need to keep leadership informed without writing essays, and want their updates actually read instead of skimmed (because every sentence moves someone toward a decision or clarity).

skill · Building unified full-stack apps when you want to avoid external backend APIs and keep everything within Next.js.

formatting-commits · When formatting-commits is needed before shipping or in production.

commit-workflow · When commit-workflow is needed before shipping or in production.

code-review-and-quality · Before merging any change when human review is unavailable.

context-engineering · When agent output quality is degrading due to insufficient or wrong context.

debugging-and-error-recovery · When something breaks and guessing wastes more time than structured triage.

documentation-and-adrs · When a significant architectural choice needs future engineers to understand the why.

git-workflow-and-versioning · When version control workflow needs to be enforced consistently.

incremental-implementation · Managing large multi-file changes by building and testing each vertical slice independently, avoiding tangled one-shot implementations.

shipping-and-launch · Shipping production changes with confidence through staged rollouts, monitoring, and tested rollback strategies.

test-driven-development · Proving code correctness through tests, making bugs reproducible and fixes verifiable.

industrial-brutalist-ui · When building data-heavy dashboards or technical portfolios that need to feel like declassified blueprints.

full-output-enforcement · When complete, unabridged output is critical and default truncation patterns would break the deliverable.

how-it-works · When applying how-it-works patterns to specialized interface challenges.

learn-codebase · When applying learn-codebase patterns to specialized interface challenges.

smart-explore · Token-efficient code navigation through symbols, functions, and dependencies—use instead of reading full files to map structure before diving into implementation.

get-unpublished-changes · Provide expert guidance at a specific stage.

ent-prompt-troubleshoot · When AI output is disappointing and the fix is in the prompt, not the model — saves time vs trying new models.

rules · Automating per-project instruction injection without rewriting CLAUDE.md for every session.

git-master · Atomic commits and complex history investigations when you need evidence-led git operations.

init-deep · Generating comprehensive, hierarchical knowledge bases for agent systems with parallel discovery.

refactor · Restructuring large code sections with safety guarantees via LSP, AST-Grep, and regression tests.

git-master · Maintaining clean commit history and enforcing conventional-commit discipline across branches.

understand-diff · Analyzing code diffs to understand what changed and its graph-based impact zone.

understand-explain · Explaining code concepts to humans using graph-derived semantic connections.

ai-native-cli · Command-line tools that AI agents will invoke without human-in-the-loop monitoring.

tdmcp-feature-discovery · When you want a prioritized feature backlog that reconciles roadmap, competitive gaps, and user ask without committing to build.

architecture-decision-records · When you need to document hard architectural decisions and keep the reasoning discoverable over project lifetime.

astro-core · Aligning rendering (SSG/SSR/hybrid + server islands) and content across an Astro site

astro-orchestrator · Dispatching Astro work to correct specialist (framework/core/wiki-publisher/creative-frontend)

carthage-env · When onboarding into a Carthage sandbox environment to understand mount scopes and permission model.

get-qodo-rules · When team coding rules must be enforced before code generation without manual rule lookup per task.

autoresearch · When running overnight experiment loops with strict keep-or-discard decisions on single-variable changes.

backend-skill · When domain-specific guidance is required.

awwwards-ui-skills · When domain-specific guidance is required.

backend-architecture-core · When domain-specific guidance is required.

crate-pr · When domain-specific guidance is required.

backend-architecture-orchestrator · Routing backend work when the specific concern (REST design, testing, deployment) is not yet named.

use-native-credential-proxy · Simple .env-based credentials for users avoiding OneCLI vault while keeping authentication straightforward.

blockchain-web3-core · Loading plugins or optional modules at runtime when keeping main bundle size small is a priority.

browser-automation-core · Deciding which browser/UI automation spoke to invoke

astropy · Astronomy research when specialized units + coordinates matter

validate-pr · Pre-merge verification when maintainer review unavailable

browser · One-shot diagnostics and deployment verification

building-native-ui · Performance-critical UIs where web is too slow

new-command · When adding new CLI subcommands to Lumen that follow register pattern, need offline operation, and must pass TypeScript type-checking.

calling-frontend-from-tauri-rust · When you need high-throughput or low-latency Rust-to-frontend communication beyond the basic event system.