By pinguinzz
Folder-as-architecture for AI agents. Replaces multi-agent frameworks with a 5-layer filesystem contract any agent can read, navigate, and extend safely. Thin guardrail skill + dispatch-subagent primitive + /icm:new (create) and /icm:assimilate (audit), backed by a canonical maintenance dept (modify-workspace · modify-room · janitor · audit) scaffolded into each workspace, a per-room tools.json model, and one stdlib checker.
Convert an existing folder into an ICM workspace, OR integrity-check an existing ICM workspace. Read-only and plan-mode — it relentlessly grills the structure, reports findings, and emits a plan for later execution. It never mutates anything.
Create new ICM structure — a workspace, department, room, or sub-room. Backbone-locked relentless intake, then scaffold into the fixed 5-layer skeleton from the plugin's templates. Additive only; never mutates existing structure.
Canonical ICM primitive — dispatch a subagent that spawns as a generic ICM agent. The subagent reads `AGENTS.md`, self-routes to a named room, finds its own data and output target, and completes the job. Use when you want work done in another room (e.g. a gated edit in the maintenance room) or a review pre-pass before requesting human review. Model (haiku/sonnet/opus) scales to the work's expected value. The dispatcher names WHERE (the room) and WHAT (the job); the subagent finds its own way.
Use when you enter a folder that follows the Interpretable Context Methodology (ICM) — signs are an `AGENTS.md` or `CONTEXT.md` at the root, numbered department/room folders (`01-`, `02.1-`), or a per-folder `docs/`. Also on "ICM", "room contract", "routing table", "sub-room", or a request to set up / audit / extend such a workspace. Loads how to enter + the base rules; points to the commands and the canon (it does not inline them).
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.
npx claudepluginhub pinguinzz/icm-plugin --plugin icmHarness-native ECC operator layer - 67 agents, 271 skills, 92 legacy command shims, reusable hooks, rules, selective install profiles, and production-ready workflows for Claude Code, Codex, OpenCode, Cursor, and related agent harnesses
Persistent file-based planning for AI coding agents. Crash-proof markdown plans (task_plan.md, findings.md, progress.md) that survive context loss and /clear, with an opt-in completion gate and multi-agent shared state. Manus-style. Works with Claude Code, Codex CLI, Cursor, Kiro, OpenCode and 60+ agents via the SKILL.md standard. Includes Arabic, German, Spanish, and Chinese (Simplified and Traditional).
Superpowers Plus core skills library for Claude Code: planning, execution routing, TDD, debugging, and collaboration workflows
Unity Development Toolkit - Expert agents for scripting/refactoring/optimization, script templates, and Agent Skills for Unity C# development
Tools to maintain and improve CLAUDE.md files - audit quality, capture session learnings, and keep project memory current.
v9.44.1 — Patch release for Gemini environment/version detection and qwen auth gating. Run /octo:setup.