orchestrator for agent skills and a library of agent skills to help with software development
This repository also configures its own agents. See what lead-producer-oss tells them →
Skill Claude Code
Frontend review team: UX quality, implementation health, accessibility, and frontend-backend integration in one recommendation. Use for UI features or frontend-heavy flows.
Skill Claude Code
Infrastructure team: deployment safety, capacity, observability, platform setup, and data-operational risk in one production-readiness assessment.
Skill Claude Code
Move/Sui team: smart contract correctness, off-chain integration, security, and economy soundness reviewed together. Use for contract changes affecting integration or economic flows.
Skill Claude Code
Open source readiness team: licensing, security, contributor experience, and strategic alignment for public collaboration.
Skill Claude Code
Product evaluation team: user value, feasibility, experience quality, and measurement in one recommendation. Use for feature proposals, go/no-go decisions, or scope negotiation.
Skill Claude Code
Adversarial review team: security, scalability, economic exploits, and behavior-based abuse analysis. Use for pre-launch or high-stakes risk analysis from attacker perspective.
Skill Claude Code
Author a new skill (role, team, or workflow) in this pack's conventions, then wire it into routing and docs. Use when adding, splitting, or formalizing a capability so it lands consistent and discoverable instead of ad hoc.
Skill Claude Code
Current-state capture workflow: orient a new owner to one bounded system, explain what it is and is not, and map the best next inspection path from code reality.
Skill Claude Code
Interface design comparison: produce 3+ competing API or module designs, compare them, and recommend one before coding. Use when multiple plausible interface shapes exist.
Skill Claude Code
Production incident workflow: detect, triage, fix, verify, postmortem. Use when something is broken in production and needs structured response with clear handoff artifacts.
Skill Claude Code
Bug handoff workflow: packages investigation findings into a durable markdown artifact with reproduction, root-cause summary, TDD fix plan, and ownership. Use when a bug needs a clean artifact for follow-up.
Skill Claude Code
Pre-mortem workflow: assume a high-stakes decision has already failed, reconstruct the failure stories backward, and turn each cause into leading indicators, mitigations, and owners before commit. Use for hard-to-reverse, launch-critical, or high-blast-radius decisions.
Skill Claude Code
Project discovery workflow: inspect an inherited or unclear repo, map major unknowns, and recommend the right timeboxed next investigation instead of guessing.
Skill Claude Code
Socratic requirements interview: grill the user about a fuzzy plan or request, force definitions, and walk the open decisions one at a time until the work is specified well enough to route. Use when a request is underspecified and starting work now would guess at intent.
Skill Claude Code
Compact the current session into a durable handoff artifact for the next session or agent: what was done, what is open, what to do next, and which skills to use. Use when wrapping up a long session or passing work to another agent.
Skill Claude Code
Narrow method for shadcn/ui work: current docs, component composition, theming discipline, accessibility.
Skill Claude Code
High-rigor review loop: run 3 contextualized specialist passes per round, score the work, and keep hardening it until the quality bar is met or a real decision blocks progress.
Skill Claude Code
Root-cause debugging workflow: investigate failures methodically, test one hypothesis at a time, and stop guessing. Use when a bug is not understood yet and the next step is investigation.
Skill Claude Code
TDD workflow adapter: routes local domain owners into strict test-first execution while preserving domain expertise and context modules.
Skill Claude Code
Project-level test suite audit, rearchitecture, and strategy. Use when the problem is the test infrastructure itself — not what to test for a single feature.
Skill Claude Code
Evidence-based elementary education review: what learning research supports or contradicts a design, how strong that evidence is, and how to measure learning outcomes. Use when a teaching approach or learning product needs grounding in research rather than intuition.
Skill Claude Code
Elementary mathematics review: learning progression, number sense, common misconceptions, concrete-to-abstract sequencing, problem difficulty, and curriculum alignment for roughly ages 6-12. Use when math content for children must be correct, well-sequenced, and teach the right idea.
Skill Claude Code
Elementary classroom review: age-appropriateness, instruction clarity, scaffolding, attention and pacing, differentiation, and feedback that motivates children roughly ages 6-12. Use when learning content or a learning product must work for real kids in a real classroom — not for subject-specific correctness.
Skill Claude Code
Icelandic language correctness gate: orthography, inflection, case government, agreement, syntax, word formation, and register, judged against current standard references. Use whenever output contains Icelandic text that must be correct — this role can block acceptance.
At most 3 mods per repository are shown here, and a mod shipped inside a plugin is left to that plugin's page — the rest are on their repository pages: