Engineering plugins

Personal engineering skills: discover, shape, specify and implement with proportionate verification.

15 skills

Engineering skills for AI coding agents, vetted and pinned: bug diagnosis, test-first work, architecture scans, plan interviews, retros, primary-source research and setup wizards. Every command runs only when you call it.

15 skills

Scheduled audits that file deduped GitHub issues, event storming and Deciders, and product modeling. Pairs with mattpocock/skills.

14 skills

14 skills: design review, KiCad verification harness, catalog maintenance, and 11 domain skills (general PCB, power, low-power, RF, audio, precision analog, high-speed, EMC/ESD, HV safety, thermal, DFM).

14 skills

Claude Code harness — JS compression hooks, per-level output rules, adversarial review gate, 4-vector compression, deny-by-default policy. Apache-2.0.

8 commands · 4 agents · hooks

Aura: a software factory of agent skills.

9 skills · 4 agents

Verified multi-model delegation: workers propose, tools prove, Claude critiques, the user authorizes.

11 skills · 1 MCP server

Daniel Noworyta's agent skills.

12 skills

Personal agent skills for real engineering.

12 skills

Route the task, explore or design before code, plan in phases with a matching loop prompt, work one slice at a time, review on the running app, hand off.

9 skills · 2 agents · hooks

Horizontal restraint layer for coding work. An always-on 'lazy senior dev' ruleset injected every turn that runs a YAGNI ladder before the agent writes code (question whether it needs to exist, stdlib before custom, native before dependency, one line before fifty), never cutting validation, error handling, security, or accessibility. Ships the referenceable `ponytail` skill plus /ponytail (lite|full|ultra|off), /ponytail-review, /ponytail-audit, /ponytail-debt, /ponytail-gain, /ponytail-help. Vendored from DietrichGebert/ponytail (MIT) v4.8.3; pairs with gstack-execution (trim then harden) and dev-infrastructure-skills.

6 skills · 6 commands

Neural Seam host adapter for the OpenAI Codex CLI. Registers the neural-seam-runtime MCP server and the $neural-seam:ns-* skills for project onboarding and the day-to-day work loop. Requires the neural-seam binary on PATH.

11 skills · 1 MCP server

Connect Codex to Cardinal telemetry and MCP.

11 skills · 1 MCP server

Cardinal Gemini CLI plugin - stream Gemini CLI telemetry to the Cardinal Outcomes Dashboard and wire the unified Cardinal MCP server.

11 skills · 1 agent

Pre-PR review and structural-smell habits built on the Foundry six-verb gate: one rule set enforced in-loop, pre-push and in CI.

10 skills · 1 MCP server · hooks

A self-routing SDLC for solo devs: grill, spec, tracer-bullet tickets, TDD build, two-axis review, retro.

11 skills · hooks

Shared subagent personas for pip-skills plugins (reviewer-correctness, -security, -scope, -rules, validator, gatekeeper and the specialists explorer, planner, implementer, verifier, summariser, researcher), built from one base with one brief/result JSON contract and its validator. Consumers declare it as a dependency; no skills.

12 agents

Use Google's Antigravity CLI from Claude Code to review code or delegate tasks.

3 skills · 7 commands · 1 agent · hooks

Route delegated work by archetype, scaffold isolated worktrees, and record each dispatch in a ledger.

2 skills · 2 commands · 6 agents · hooks

Work Matt Pocock's way, ship OpenSpec artifacts: grill-me interviews feed OpenSpec change proposals, and implement/tdd drive OpenSpec task lists.

6 skills · 4 commands · hooks

Eight-skill methodology for architecture investigations: grill the user until scope is concrete, write a one-page brief, draft a junior-followable plan with a per-item output schema, execute per-item in batched subagents with per-step verification and peer review, then synthesize a recommendation and close with a head-of-architecture review.

10 skills · 1 command

Uzasch's agent skills — implement-and-review loops with a pluggable reviewer (Codex, Claude, or Google Antigravity), and more to come.

11 skills

Coordinate a work track, brief lanes, review and gate PRs, write PR bodies, promote a release, untangle a codebase, choose a stack, audit a lane's decisions, eval a skill blind, deepen a task before starting it.

11 skills

Ten advisory skills covering the six stages of the AI-native SDLC: intent, spec, plan, CLAUDE.md, feedback loop, evals, PR review, approval gates, and closing the loop, plus a project audit advisor.

10 skills

Skills for each phase of our workflow (setup-project, research, plan-milestones, work-on-issue, review-pr) plus our shared agents.

5 skills · 5 agents

コード開発者向けの規約集（mise、hk、GitHub Actions、Git の設定、Python、PowerShell）を、AI エージェントが従う形で定める。あわせて、GitHub 初学者に Git、Issue、Pull Request の使い方と、リポジトリの初期設定を案内する。

10 skills

Security engineer — threat modelling, security audits, compliance, vulnerability management

7 skills · 2 agents

Humphrey Ahn's agent skills: writing code a human can read a month later, and building skills that hold up.

9 skills

Persistent work ledger plus card orchestration: delegated agents, adversarial review loops, budget tripwires and session handoff.

2 skills · 1 command · 5 agents · hooks

Nine skills for agentic engineering: doctor, compass, cartographer, crucible, council, reproduce, proof, red-team, relay.

9 skills