The library
Everything we index — ranked by what works, never by stars.
forSalesMarketingHRFinanceLegalOpsProductEngineeringDataProductivitySupportsetup≤ plug & play≤ + a key≤ multi-tool
● works · ● untested / no effect · ● hurts — every rank is measured against a no-skill baseline
untested★17,464→untested★0→untested★9→untested★1→untested★1→untested★0→untested★0→untested★204→untested★360→untested★439→untested★294→untested★80→untested★155→untested★158→untested★92→untested★127→untested★97→untested★35→untested★39→untested★34→untested★41→untested★33→untested★40→untested★14→untested★7→untested★8→untested★25→untested★16→untested★16→untested★30→untested★8→untested★6→untested★4→untested★10→untested★27→untested★7→untested★7→untested★13→untested★6→untested★3→untested★3→untested★12→untested★6→untested★5→untested★11→untested★2→untested★3→untested★3→untested★6→untested★0→
Update all documentationcommandEngineeringL1
update-docs · Post-creation workflow ensuring new skills reach all platforms (Codex, Gemini, OpenClaw), marketplace stays current, and docs auto-generate.
Research priority ticketscommandEngineeringL1
ralph_research · Systematic issue research workflow: prioritize, investigate codebase and external sources, synthesize findings, attach to issue and move to review.
Find code vulnerability patternscommandEngineeringL2
contracts:audit-variants · Security audit: find and classify variants of a known vulnerability pattern across the entire repo, separating confirmed/probable/false-positive.
Configure safety guardrails for developmentcommandEngineeringL1
harness · Prevent infinite loops and unwanted file changes in autonomous Claude Code sessions via iteration caps, timeouts, protected paths, and required checks.
Consult expert LLM council on decisionscommandL1
council · Soliciting diverse perspectives on architecture/tech/implementation decisions by running a query through multiple LLMs and synthesizing a consensus.
Capture task ideas as todoscommandProductivityL1
gsd:add-todo · Context-preserving task capture during active work: record idea/task without derailing current session, with enough detail for future Claude to understand.
Run focused discussion to reach conclusioncommandL2
{"Spectra":"Discuss","description":"Have a focused discussion about a topic and reach a conclusion","category":"Workflow","tags":["workflow","discuss","thinking"]} · Structured design discussion that narrows options, surfaces key trade-offs, makes a recommendation, and converges on a decision before implementation.
Critique PR review for accuracycommandEngineeringL1
review-the-review · Secondary review gate: catch sycophancy, false positives, and missed issues in bot reviews; converge on approve/request-changes/defer via up to 10 rounds.
Generate git commit messagescommandEngineeringL1
func-commit-message · Commit message generation: analyzes staged changes and drafts a message following conventional commit format, seven rules, and imperative mood.
Optimize CI workflow performancecommandEngineeringL2
optimize · CI performance tuning: measure baseline, identify bottleneck steps, detect common issues (concurrency, caching, runner size), recommend Blacksmith optimizations.
Query knowledge base for answerscommandL2
wiki-query · Knowledge synthesis: find relevant wiki pages, synthesize answer with [[wikilink]] citations, optionally save synthesis page for reuse.
Enter document-driven development modecommandEngineeringL1
ddd · Multi-phase development workflow: start with requirements, move through specs/plan/implementation/documentation, with explicit approval gates between phases.
Start local development environmentcommandEngineeringL2
dev · Quick local dev startup: orchestrates Docker middleware, FastAPI backend, dual Vue frontends (admin + web) on unified Nginx gateway.
Handle git tasks with conventionscommandEngineeringL1
git · Enforced git workflow: branch names (type/scope-description), commits (conventional format), avoids destructive ops, detects and prevents secret commits.
Audit registry for violationscommandEngineeringL2
registry-audit · Read-only component registry verification: checks registry/filesystem alignment, schema validity, dependency validity, and circularity.
Fix BTC E2E test failurescommandEngineeringL2
fix-btc-e2e-p5 · Native SegWit single-sig BTC test repair: targets bech32 addresses, wpkh descriptors, Pattern 5 workflows.
Run pre-release checklistcommandEngineeringL1
release-checklist · Structured release process: pre-release validation, tests/security/docs verification, staging tests, production deployment, post-deploy monitoring, rollback plan.
Generate X threads from contentcommandMarketingL1
publisher-x · Turning long-form content into multi-format social media threads with minimal effort
Review strategic plans with CEO lenscommandOpsL1
gstack-plan-ceo-review · CEO/founder-level plan review where scope control and ambition calibration matter
Check code for AI agent patternscommandEngineeringL1
vibe-check · Pre-commit quality check when AI agents have contributed code
Start guided learning sessioncommandProductivityL1
analysis-learn · Structured onboarding to data analysis workflows with feedback loops and skill progression
Audit pages for SEO issuescommandMarketingL2
web:seo-audit · Pre-launch SEO validation when site structure and metadata tuning are critical
Validate decisions with consensus panelcommandL1
validate-consensus · High-stakes design decisions where independent expert review reduces blind spots
Find performance bottlenecks fastcommandEngineeringL3
perf · Full-stack mobile perf audit when you need to find unbounded queries and memory leaks simultaneously
Execute tasks with window optimizationcommandL1
local.implement · Multi-phase implementation where context window and task state must be preserved across sessions
Prepare new framework releasecommandEngineeringL1
release · Multi-file version synchronization when human approval before commit is required
Draft pull request descriptioncommandEngineeringL1
git-pr-description · Structuring PR descriptions when coverage of every commit is non-negotiable
Install skill domain into projectcommandEngineeringL2
install-domain · Bulk-deploying skill packages across multiple agent projects
Install skill bundle into projectcommandEngineeringL1
install-bundle · Quick onboarding of new projects with a comprehensive skill set
Promote priority issues for reviewcommandOpsL1
champion-issue-promo · Three-tier approval workflow ensuring only goal-aligned issues reach the backlog
Cluster instincts into reusable skillscommandEngineeringL2
evolve · Continuous code quality improvement when you need structured feedback on engineering patterns
Run spaced-review for knowledge vaultcommandProductivityL1
recall · Retrieving domain expertise without cloud egress when privacy is critical
Run full pre-commit check suitecommandEngineeringL1
pre-commit · Preventing common mistakes from reaching version control
Identify and fix security issuescommandEngineeringL1
security-review · Pre-release security scanning when compliance and threat coverage matter
Extract design system valuescommandProductL1
devlyn.design-system · Single-source-of-truth design system when consistency across products is required
Run all pre-merge gates in parallelcommandEngineeringL1
release-check · Preventing incomplete releases through automated safety checks
Check diff coverage against CI gatescommandEngineeringL1
coverage-diff · Preventing coverage regressions when test quality gates are enforced
Connect landing page to email funnelcommandMarketingL3
workshop-marketing:pagina-active · Connecting high-conversion capture pages to email automation without backend overhead or Zapier friction.
Broadcast message to all agentscommandOpsL2
adjutant.broadcast · Coordinating standing orders or announcements across a live swarm of distributed agents in real-time.
Capture and monitor LSP logscommandEngineeringL2
watchlogs · Debugging LSP editor integration when client-side requests (completions, hover) behave unexpectedly.
Apply OpenSpec changes to codebasecommandEngineeringL2
OPSX: Apply · Executing spec-driven implementation with fluid artifact refinement and enforced task tracking.
Create GitHub issue via CLIcommandEngineeringL1
create-issue · Creating GitHub issues programmatically without leaving the CLI or entering the web UI.
Scout, plan, build with agile trackingcommandEngineeringL3
scout_plan_build_v2 · Full-cycle agile feature delivery: requirements → architecture → Confluence docs → Jira tasks → implementation → metrics.
Improve UI with design agentcommandProductL2
improve-ui · Systematically redesigning components against a strict design system to prevent visual drift and generic LLM output.
Copy-paste workflow command sequencescommandOpsL1
workflow-examples · Running standardized NotebookLM research workflows (trends, SWOT, learning, deep-dive) with repeatable output.
Save conversation session as reportcommandProductivityL1
save-session · Archiving Claude Code conversation sessions as timestamped reports for async team handoff or retrospective.
Create isolated git worktrees safelycommandEngineeringL1
sh-worktree · Isolating feature branches with automatic setup and baseline test verification to prevent cross-branch contamination.
Estimate development effort with sizingcommandOpsL1
sh-estimate · Right-sizing task effort with risk multipliers and uncertainty quantification for sprint planning.
Fuse specs and autotuned baselinescommandEngineeringL3
phase2-implement · Implementing autotuned kernel baselines from validated specifications before Phase 3 algorithmic exploration.
Start development servercommandEngineeringL1
run · Domain-specific command workflow.