Enforce evidence-gated development workflows that prevent self-asserted 'done' from passing — re-derives results through independently attested gates, traces code relationships beyond static analysis, runs multi-model second opinions, and applies structured archetypes (build, migrate, ship, review, incident response) with rigor scaling to complexity and blast radius.
Deep trust and safety audit for agentic systems with risk classification and compliance validation
Interactive agentic architecture design guidance with pattern recommendations and safety validation
Use when selecting or comparing agentic frameworks (LangChain, CrewAI, AutoGen, etc.) for a specific use case — gets latest ecosystem context via WebSearch. NOT for reviewing existing agentic code (use agentic:review) or architecture design (use agentic:design).
Full agentic codebase review with framework detection, agent topology analysis, and remediation roadmap
Implement and ship a feature or fix — test rigor scales with signals
Five-layer architecture validation, agent topology analysis, orchestration... Use when: architecture review, agent...
Token optimization, latency budgets, cost analysis, parallelization... Use when: performance optimization, cost...
Guardrails, prompt injection defense, PII protection, human-in-the-loop gates,... Use when: safety review,...
Execute implementation tasks according to approved designs and test strategies. Use when: building features,...
Explore codebase and gather context before design or implementation choices. Use when: codebase exploration, pattern...
v11 work-shape archetype runner. When a prompt has been routed to one of the archetypes (triage, explore, specify, decide, ship, review, incident, build, migrate, modernize), this skill is the entry point. It picks the right per-archetype playbook from refs/ and executes the phase shape declared in `.claude-plugin/archetypes.json`. Use when: a `<wg archetype="X">` or `<wg archetypes>` system-reminder tag appears, an explicit "let's run the X archetype" request, or when one of the per-archetype slash commands resolves to this skill.
v11 LLM-based work-shape classifier. Replaces the regex archetype detector with the model's own reasoning. Reads the user's prompt, picks the right archetype(s) from the catalog, identifies signals (blast_radius, novelty, reversibility, etc.), and persists to SessionState so subsequent turns steer correctly. Use when: the prompt_submit hook emitted a `<wg classify-due />` directive, OR explicitly invoked at session start, OR when re-classifying after the user changes scope mid-session.
Critical thinking framework applied before doing work. Challenges assumptions, reframes problems, identifies hidden opportunities, and validates whether the stated ask is the right ask. A way of approaching work, not a specialist discipline. Use when: challenging assumptions before implementation, reframing a stated problem, or asking whether the right thing is being asked at all.
Grounding / assumption-check: pull deeper context from brain + bus when uncertain, before acting. Use when: getting mixed signals from the codebase, about to commit to a non-obvious decision, prior decisions might exist for this exact problem, or you want to verify an assumption before action ("am I sure about this?", "have we decided this before?"). Returns relevant brain memories, recent bus events, and linked priors ranked by relevance — not a wall of text. Aliases: grounding, sanity-check, verify-assumption. NOT for: routine "what does this code do" questions (use Read or Grep), broad codebase exploration (use Agent(Explore)), or fetching specific symbols (use wicked-brain:search directly).
Capability router that decides which tools, skills, and agents to use for a task. Discovers CLI tools in PATH alongside MCP servers, skills, and agents. Reduces cognitive load on the main agent by making tool selection decisions. Use when: planning work in an unfamiliar domain, deciding which tools or agents fit a task, or discovering installed CLIs / MCP servers.
Matches all tools
Hooks run on every tool call, not just specific ones
Executes bash commands
Hook triggers when Bash tool is used
Own this plugin?
Verify ownership to unlock analytics, metadata editing, and a verified badge. GitHub access is read-only (username + org membership).
Sign in to claimOwn this plugin?
Verify ownership to unlock analytics, metadata editing, and a verified badge. GitHub access is read-only (username + org membership).
Sign in to claimBased on adoption, maintenance, documentation, and repository signals. Not a security audit or endorsement.
Modifies files
Hook triggers on file write and edit operations
Modifies files
Hook triggers on file write and edit operations
Uses power tools
Uses Bash, Write, or Edit tools
Uses power tools
Uses Bash, Write, or Edit tools
_ _ _ _
__ _(_) ___| | _____ __| | __ _ __ _ _ __ __| | ___ _ __
\ \ /\ / / |/ __| |/ / _ \/ _` |_____ / _` |/ _` | '__/ _` |/ _ \ '_ \
\ V V /| | (__| < __/ (_| |_____| (_| | (_| | | | (_| | __/ | | |
\_/\_/ |_|\___|_|\_\___|\__,_| \__, |\__,_|_| \__,_|\___|_| |_|
|___/
Your coding agent already plans and swarms. Wicked Garden is the curated toolkit for what it can't do alone.
📖 Docs → mikeparcewski.github.io/wicked-garden. Identity & beliefs →
ETHOS.md. How it works →CLAUDE.md.
Coding agents grew up. Claude Code, Codex, Cursor, Antigravity, Aider, OpenCode, Gemini CLI, Zed/ACP — they're not autocomplete anymore. They plan. They parallelize. And each has strong opinions about how it likes to work.
Most plugins try to boss them around — re-implement planning, impose a workflow, make the agent dance. You end up fighting your own tools.
Wicked Garden refuses to wrestle the harness. It assumes your agent is good at the things it's good at, and fills the gaps it can't fill on its own.
| Your harness… | Wicked Garden… |
|---|---|
| says "tests pass" (sometimes it's lying) | re-runs the proof. False "done" → rejected. Missing backend → fails closed. Never a vacuous green. |
| greps and reads — blind to string-wired links | sees the injected edges (event→consumer, command→agent, agent→capability) grep never will → blast-radius, lineage |
| refactors on a hope and a prayer | renames across files as a graph operation, not find-replace roulette → wicked-patch |
forgets everything at exit | remembers what session 1 decided when you're in session 47 → wicked-brain |
| re-derives how to work in this repo every task — which file owns the bug, the wiring step, the test command | loads the repo's own playbooks (fix-bug/add-feature/verify…), generated from HEAD → wicked-understanding |
| asks itself for a second opinion | convenes a real multi-model panel (Gemini / Codex / …) → jam:council |
| re-derives WCAG/CWE/SOC2 from memory every time | loads the rubric on demand, ships it to any repo |
| grades its own homework | author ≠ executor ≠ reviewer → evidence-gated testing |
The throughline: done is re-derived, not asserted. Verdicts you can trust on the first read — not green checkmarks you can't.
/wicked-garden:compile). Stand on the harness, fill its gaps, hand off. Never absorb.No universal pipeline to obey. A hook reads each prompt's shape and that decides one thing: how much rigor this work earns. A typo (triage) gets none; a migration cutover (migrate) gets a hard, independently-attested gate with a rollback proof. Nine shapes — triage · explore · specify · decide · build · review · ship · incident · migrate — steering, not blocking. Why shapes and not one pipeline → docs/v11/archetypes.md.
/plugin install wicked-garden # in Claude Code
/wicked-garden:setup # verifies peers; blocks only on the two the gate needs
Two required, three opt-in. The evidence gate is the floor we won't fake, so it needs two peers — setup blocks without them:
npx wicked-vault-install # the honest-evidence backend the gate re-derives against (≥ 0.3)
npm i -g wicked-loom # the gate engine that re-runs the verifier (or via npx)
The rest of the kit is opt-in layers — add what you want, skip the rest and the toolkit still works:
npx wicked-testing install # the acceptance-testing layer (author ≠ executor ≠ reviewer)
/plugin install wicked-brain # the memory layer — cross-session recall + cited search (the "what")
npx skills add mikeparcewski/wicked-understanding --all # the repo-playbooks layer — "how to work in THIS repo" (the "how"; pairs with brain)
/plugin install wicked-bus # the audit-trail layer (fire-and-forget; fail-open without it)
Optional, lights up the code graph: npx @colbymchenry/codegraph (Node ≥ 22.5) → powers blast-radius / lineage / hotspots / wicked-patch. Details: docs/required-peers.md.
npx claudepluginhub mikeparcewski/wicked-garden --plugin wicked-gardenRegisters the wicked-estate MCP server: a code + infrastructure estate graph for agents — definitions, who-calls-X, blast-radius, scoped context, across 91 languages plus mainframe/IaC estate.
Interactive HTML & presentation builder with an in-browser feedback loop for non-technical business users. Build a draft, review it in the browser, highlight any block, attach plain-language feedback, and watch it regenerate live — with the supervising Claude session as the intelligence in the loop. Versioned, forkable, and exportable to self-contained HTML or PDF.
Standalone QE library — 5-core testing surface (plan/authoring/execution/review/insight), SQLite ledger with fixed-SQL oracle, optional wicked-bus + wicked-brain integration.
Comprehensive skill pack with 66 specialized skills for full-stack developers: 12 language experts (Python, TypeScript, Go, Rust, C++, Swift, Kotlin, C#, PHP, Java, SQL, JavaScript), 10 backend frameworks, 6 frontend/mobile, plus infrastructure, DevOps, security, and testing. Features progressive disclosure architecture for 50% faster loading.
Comprehensive feature development workflow with specialized agents for codebase exploration, architecture design, and quality review
Develop, test, build, and deploy Godot 4.x games with Claude Code. Includes GdUnit4 testing, web/desktop exports, CI/CD pipelines, and deployment to Vercel/GitHub Pages/itch.io.
Comprehensive PR review agents specializing in comments, tests, error handling, type design, code quality, and code simplification
Intelligent draw.io diagramming plugin with AI-powered diagram generation, multi-platform embedding (GitHub, Confluence, Azure DevOps, Notion, Teams, Harness), conditional formatting, live data binding, and MCP server integration for programmatic diagram creation and management.
Access thousands of AI prompts and skills directly in your AI coding assistant. Search prompts, discover skills, save your own, and improve prompts with AI.