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★0→untested★1→untested★2→untested★0→untested★0→untested★2→untested★24→untested★24→untested★0→untested★1→untested★0→untested★2→untested★0→untested★1→untested★0→untested★0→untested★2→untested★0→untested★0→untested★0→untested★1→untested★2→untested★1→untested★1→untested★0→untested★1→untested★0→untested★0→untested★0→untested★1→untested★0→untested★0→untested★0→untested★0→untested★1→untested★0→untested★0→untested★0→untested★0→untested★1→untested★0→untested★1→untested★0→untested★0→untested★0→untested★1,619→untested★289→untested★227→untested★241→untested★358→
Scaffold architecture decision recordcommandEngineeringL1
new-adr · Automating new-adr workflow in t11z/kenny with validation gates and structured output.
Release versioned product updatecommandEngineeringL1
release · Publishing releases to Flux-Frontiers/pycode_kg with automated versioning and changelog updates.
Capture design patterns to memorycommandEngineeringProductL1
capture · Documenting design tokens and decisions in lfades/postcards for future reference.
Initialize new BaseNative projectcommandEngineeringL2
bn-init · Bootstrapping BaseNative projects with standard folder structure and config.
Execute WarmDesk releasecommandEngineeringL1
release · Publishing releases to tonk/warmdesk with automated versioning and changelog updates.
Build and validate UI paritycommandEngineeringL2
fabric-parity-loop · Validating UI parity between Fabric and custom Loom implementations at scale.
Parse GTM strategy argumentscommandL1
dev-gtm · Quickly route different developer documentation audits to the right specialized tool without manual dispatch.
Parse marketing campaign argumentscommandL1
marketing · Dispatch marketing tasks across 20+ specialized skills without memorizing each command syntax.
Run interactive fine-tune pipelinecommandEngineeringL2
finetune · Iteratively improve fine-tuned models with human approval gates at each stage, avoiding costly training mistakes.
Create NextDocs documentationcommandEngineeringL1
nextdocs · Quickly bootstrap a documentation site structure that follows your org's NextDocs convention without manual boilerplate.
Preflight code and propose RIAL commitcommandEngineeringL2
rial-ship · Ensure code passes all quality gates (type check, lint, test, build) before proposing a release commit.
Scaffold and validate Pulse page speccommandEngineeringL2
build-page · Quickly scaffold Pulse pages with visual validation before running tests, preventing spec-to-render failures.
Manage active impediment statecommandProductivityL1
impediment · Track what's blocking current focus without polluting the objective layer, surfaces in every prompt's banner.
Explore tree search with MCTS evaluationcommandEngineeringL2
search-tree · When a question has multiple valid approaches and you want honest trade-off analysis before deciding which path to take.
Review GitHub PR and post commentscommandEngineeringL3
pr-reviewer · Post reviewed comments to PRs with human approval checkpoint, avoiding accidental toxic feedback or wrong decisions.
Conduct full post-implementation code reviewcommandEngineeringL2
review-code · Verify post-implementation code complies with FOAA architecture (dependency direction, module structure, security) without guessing.
Lint Ansible configuration filescommandOpsL1
ansible-lint · Dispatch commands to the right handler without boilerplate.
Evaluate resume against job descriptioncommandHRL1
judge-resume · Dispatch commands to the right handler without boilerplate.
Run vaultlab result analysis pipelinecommandDataL2
run-analysis · Dispatch commands to the right handler without boilerplate.
Route user request to best workflowcommandOpsL2
run-workflow · Dispatch commands to the right handler without boilerplate.
Generate technical documentation from speccommandEngineeringL1
doc-write · Dispatch commands to the right handler without boilerplate.
Create SEO-optimized blog postcommandMarketingL1
new-blog-post · Dispatch commands to the right handler without boilerplate.
Launch parallel agents on epic taskscommandOpsL3
epic-start · Dispatch commands to the right handler without boilerplate.
Setup macOS development environmentcommandOpsL2
new-macos-setup · Dispatch commands to the right handler without boilerplate.
Diagnose platform issuescommandOpsL2
debug · Dispatch commands to the right handler without boilerplate.
Orient on latest branch and propose actioncommandProductivityL1
pickup · Resuming work across devices when session context is unclear
Audit documentation quality comprehensivelycommandEngineeringL2
[FRG]-docs-audit · Identifying documentation debt and stale docs before they impact new contributors
Backup and restore configurationcommandOpsL1
backup · Restoring application state and configuration after a crash or when migrating devices
Decode ERC-8021 dataSuffix hexcommandEngineeringL1
decode · Debugging ERC-8021 builder codes in calldata without an RPC roundtrip
Auto-compute rote calculation answerscommandProductivityL1
crunch · Escaping rote algebra after conceptual understanding is locked in via /learn
Add skill entry to React componentcommandEngineeringL1
add-skill · Incrementally building a curated skills portfolio without manual array formatting
Score skill quality with paired checklistcommandEngineeringL1
evaluate-skill · Rapid quality assessment of buddy skills before shipping
Run pre-deployment checkscommandOpsEngineeringL1
deploy · Preventing silent production bugs by enforcing code quality and config checks before deploy
Specify screens like a senior designercommandProductEngineeringL1
design · Converting vague user stories into implementation-ready design specs with no designer follow-up
Loop superintendent task on schedulecommandOpsL1
hermit · Orchestrating parallel work on multiple independent issues without waiting for serial completion
Verify multi-path formulas end-to-endcommandEngineeringDataL1
verify · Catching convention mismatches and bulk-boundary inconsistencies in advanced mathematics
Fix bugs fast while maintaining standardscommandEngineeringL1
devspark.quickfix · Shipping rapid hotfixes while staying aligned with project constitution
Implement technical plans with PR flowcommandEngineeringL2
implement_plan · Executing multi-phase technical implementations with built-in phase verification and pause points
Report and track bugs systematicallycommandEngineeringProductL1
fix-bug · Rapidly fixing startup/tooling bugs without broad code exploration
Map data flows and debug end-to-endcommandEngineeringL1
trace · Finding where a feature or bug lives by tracing the complete data flow with file:line accuracy
Review and fix documentationcommandEngineeringOpsL1
docs · Keeping documentation in sync with code changes across the entire project
Author a new agent skillcommandEngineeringL1
create-skill · Authoring new agent skills that follow Cursor conventions without manual boilerplate
Load project context at session startcommandEngineeringL1
bootstrap · Starting work in a new session with full context awareness
Create and prioritize roadmapscommandProductOpsL1
doit.roadmapit · Building prioritized product roadmaps aligned with constitution principles
Summarize session issues and fixescommandEngineeringOpsL1
summary · Capturing session work products (code reviews, fixes, findings) immediately after completion without manual re-scanning.
Cut releases across all platformscommandEngineeringOpsL2
release · Multi-platform Debian-based software releases requiring coordinated PPA, PyPI, Homebrew, and GitHub synchronization with RC iteration.
Fix grammar in tracked filescommandL1
fix-grammar · Light grammar-only fixes on blog posts before publication; simple file-level corrections without structural changes.
Guide test-driven developmentcommandEngineeringL1
tdd · Tasks requiring disciplined test-first development with explicit Red-Green-Refactor cycles and framework-agnostic guidance.
Resume session with pruned contextcommandL1
prune-resume · Extending long-running Claude Code sessions by pruning token-heavy conversation history while preserving session context and resuming automatically.
Start lesson 18.2 in EnglishcommandL1
start-18-2.en · Structured product management workflow where requirements must be extracted from customer research and prioritized using MoSCoW method within a single lesson session.