The library

Everything we index — ranked by what works, never by stars.

untested
Manage autonomous X/Twitter presencesubagentMarketingL3
x9 · Multiplayer games that need consistent player-state synchronization with latency compensation and conflict resolution.
untested
Coordinate financial team strategysubagentFinanceL3
max · Orchestrating TradingIA development when someone needs integrated strategy across code, math, market, and business.
untested
Design and write database schemasubagentEngineeringL1
dba · When data layer must be explicit and performant before backend implements repositories.
untested
Analyze unit economics and marginssubagentFinanceL1
The CFO · Reality-checking startup assumptions when the board needs to hear if numbers actually work.
untested
Evaluate technology and hosting choicessubagentEngineeringL1
cto · Making infrastructure or agent-backbone choices with cross-cutting impact across teams.
untested
Build RAG and knowledge systemssubagentEngineeringL1
rag · Building semantic search over regulatory documents (Verordnungen) when 700-word chunks matter.
untested
Optimize database schema and queriessubagentEngineeringL1
dba · Building semantic search over regulatory documents (Verordnungen) when 700-word chunks matter.
untested
Manage projects and dispatch worksubagentOpsL1
pam · Front-door assistant when requests mix personal advice, capture, and project execution.
untested
Generate SEO tags and structured datasubagentMarketingL1
seo · Improving search ranking when structured data and crawlability are critical.
untested
Write production Python codesubagentEngineeringL1
sde · Staff-level coding when production discipline and maintainability matter over speed.
untested
Investigate PR history and repository statesubagentEngineeringL1
gitnexus-pr-facts-historian · Understanding PR scope before review when changed files, commits, and regressions matter.
untested
Conduct authorized security penetration testssubagentEngineeringL2
penetration-tester · Demonstrating real security risk when authorized exploitation and impact assessment needed.
untested
Guide architecture and codebase decisionssubagentEngineeringL1
edgartools-architect · Planning major EdgarTools changes when beginner-friendly simplicity must survive refactor.
untested
Onboard models to qai_hub_modelssubagentEngineeringL2
onboarding · Integrating edge models when monkeypatching and compilation workarounds need expert guidance.
untested
Solve CTF and hacking challengessubagentEngineeringL2
ctf-solver · Solving web/pwn/crypto CTFs when systematic methodology beats guessing.
untested
Bump tool and dependency versionssubagentEngineeringL1
version-bumper · Automating releases when version coordination across repos needed.
untested
Apply specified code fixessubagentEngineeringL1
pr-fix-implementer · Executing code fixes when the design is locked but implementation needs care.
untested
Create new database entity classessubagentEngineeringL1
new-entity · Creating ORM entities when schema design needs code generation.
untested
Simplify code without changing behaviorsubagentEngineeringL1
code-simplifier · Cleaning code when reuse wins matter more than bug hunting.
untested
Update and sync documentationsubagentEngineeringL1
docs-agent · Keeping docs fresh when code changes need user-facing explanation.
untested
Research technical solutions onlinesubagentEngineeringL2
ark-researcher · Analyzing agent-at-scale patterns when McKinsey research needs synthesis.
untested
Manage Uniswap v3 liquidity positionssubagentFinanceL3
{"{ projectName }":null} · Autonomous LP range management when automated rebalancing within USD bounds beats manual monitoring
untested
Detect bugs in pull request codesubagentEngineeringL1
pr-reviewer-bugs · Catching null-check gaps and breaking changes when deep code reading and cross-file validation matter
untested
Create new host configurationsubagentEngineeringL2
gstack-contrib-add-host · Adding agent support to gstack when a new coding assistant needs skill reuse across host ecosystems
untested
Verify documentation accuracysubagentEngineeringL1
documentation-accuracy-reviewer · Ensuring user docs match implementation when multi-layered documentation must stay in sync
untested
Normalize React and TypeScript codesubagentEngineeringL2
react-normalizer · Applying company standards consistently when 2-iteration cap and business-logic preservation matter
untested
Analyze weather patterns and trendssubagentDataL2
weather-analyst · Multi-period weather analysis when hour-by-hour comparisons and extreme event detection beat simple forecasts
untested
Map sensitive data flows for securitysubagentEngineeringL3
data-flow-tracer · Security audit when complete data-flow visualization and validation-gap discovery outweigh endpoint lists
untested
Explore read-only vault documentationsubagentEngineeringL2
explorer · Non-mutating intelligence gathering when read-only analysis and pattern detection matter
untested
Improve structured logging and observabilitysubagentEngineeringL3
logging-subagent · High-concurrency Rust when structured field logging and span-based context tracing beat free-text logs
untested
Maintain AI instructions and agent architecturesubagentEngineeringL3
vibe-architect · Multi-skill governance when durable rule files and centralized routing beat scattered configs
untested
Conduct deep technical auditssubagentEngineeringL3
reasoner · Honest technical assessment when architectural health and productization readiness matter
untested
Manage events lifecycle and sponsorshipssubagentMarketingL2
events-agent · Full event coordination when checklist-driven lifecycle beats fragmented spreadsheets
untested
Review econometric identification strategysubagentDataL3
causal-inference · Causal inference when observational data and rigorous assumption testing beat correlation studies
untested
Analyze codebase and generate intelsubagentEngineeringL3
gsd-intel-updater · Intelligence database maintenance when automated updates and validation matter
untested
Write idiomatic C++ with modern patternssubagentEngineeringL3
cpp-pro · C++ performance tuning when measured improvements and correctness verification matter
untested
Answer mid-session repo questionssubagentEngineeringL2
qa-agent · QA automation when systematic test execution and coverage metrics beat ad-hoc testing
untested
Triage backlog and prioritize worksubagentProductL2
pm · Project coordination when centralized status tracking and risk visualization matter
untested
Review Rust code for safety and idiomssubagentEngineeringL2
code-reviewer · Catching Rust safety bugs in compiler code reviews without full AST context.
untested
Develop Flutter mobile applicationssubagentEngineeringL3
flutter-expert · Mentoring a team on Flutter patterns when Riverpod state is complex.
untested
Advise on SaaS tax and accountingsubagentFinanceL2
felix · Configuring entity structure and tax planning at SaaS hire-time.
untested
Audit and generate tests across repossubagentEngineeringL3
README · Parallel auditing of unfamiliar legacy code when multiple specialists are needed.
untested
Organize and triage note inboxsubagentProductivityL2
sorter · Batching daily Obsidian inbox triage to maintain project momentum.
untested
Audit SEO health and crawl sitesubagentMarketingL3
seo-auditor · Detecting SEO regressions after major Vercel deploys without manual crawling.
untested
Refactor Rails code after tests passsubagentEngineeringL2
rails-refactoring-expert · Cleaning up working Rails code while keeping tests green.
untested
Formalize plans into specificationssubagentProductL2
architect · Turning approved plans into precise specs before dev teams start coding.
untested
Write clear documentation and ADRssubagentEngineeringL2
documenter · Writing consistent architecture docs when live code outpaces documentation.
untested
Review code quality and securitysubagentEngineeringL2
quality-reviewer · Universal PR quality checks when no role-specific reviewer exists.
untested
Review technical designs and feature planssubagentEngineeringL3
sqlbuildingblocks-plan-reviewer · Stress-testing feature plans against dialect complexity and NonTerminal dispatch.
untested
Write and run tests for leto-rssubagentEngineeringL2
leto-tester · Applying deep expertise to a focused scope.
page 146 / 156