By ryan-mt
Strict 3-role orchestration plugin — Orchestrator plans, Workers code, Validators verify. Hard enforcement via PreToolUse hooks.
Enter Mission Mode — strict 3-role orchestration (Orchestrator, Worker, Validator)
Exit Mission Mode — emergency stop that always works, bypasses all guards
View or set Mission Mode configuration — models for orchestrator, worker, validator, persistence, and flags
Show current Mission Mode progress — phase, round, per-feature status from features.json
Mission Orchestrator — plan, delegate, review. NEVER writes code. Used as a reference for the main session when in Orchestrator phase.
Mission Worker — implements code. NEVER plans or validates.
Mission Validator — verify, test, break things. NEVER writes production code.
Executes bash commands
Hook triggers when Bash tool is used
Modifies files
Hook triggers on file write and edit operations
Uses power tools
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.
Uses Bash, Write, or Edit tools
Uses Bash, Write, or Edit tools
A strict 3-role orchestration system that turns Claude Code into a disciplined engineering team. The Orchestrator plans, Workers code, and the Validator tests — each role isolated by hard hook enforcement at the tool-call level. Features structured tracking via features.json, model enforcement, and Factory-grade anti-drift protection.
Claude Code is powerful, but without structure it can skip tests, mix concerns, or lose track of quality. Mission Mode forces a disciplined loop:
Plan → Create features.json → Dispatch Workers per feature → Validate → Fix → Validate → ... → Done
No role can do another role's job. Workers can't run tests. Validators can't write source code. The Orchestrator can't touch files outside .mission/. All enforced by Python PreToolUse hooks — not just prompts.
# From GitHub
/plugin marketplace add riftzen-bit/mission
/plugin install mission
# Or from a local clone
/plugin install /path/to/mission
/enter-mission Build a REST API for user authentication
That's it. Mission Mode activates and runs the full loop automatically:
features.json with structured feature list| Command | Description |
|---|---|
/enter-mission [task] | Start Mission Mode (or resume if interrupted) |
/exit-mission | Emergency stop — always works, bypasses all guards via endedAt |
/mission-status | Dashboard: phase, round, per-feature progress from features.json |
/mission-config | View or set models for orchestrator, worker, validator + persistence |
Plans, delegates, reviews. Spawns Workers and Validators via the Agent tool.
.mission/, create/update features.json, spawn agentsImplements code exactly as assigned per feature from features.json. Produces structured JSON handoffs.
.mission/state.json, spawn ValidatorsThe ruthless quality gate. Validates per-feature, writes tests, runs all checks, finds bugs.
.mission/reports/*, run any commandRoles are enforced at the tool-call level by a Python hook engine (hooks/engine.py) shared across eight hooks:
hooks/phase-guard.py (PreToolUse) — Blocks forbidden tool calls per phasehooks/mission-reminder.py (PreToolUse) — Injects role-specific anti-drift reminders with feature contexthooks/mission-continue.py (PostToolUse) — Injects continuation reminders with strength gradienthooks/mission-stop.py (Stop) — Blocks session from ending while mission is activehooks/mission-subagent-stop.py (SubagentStop) — Blocks sub-agents from premature stoppinghooks/mission-precompact.py (PreCompact) — Saves checkpoint before context compactionhooks/mission-session-start.py (SessionStart) — Auto-injects mission context on session start/resumehooks/mission-prompt.py (UserPromptSubmit) — Injects mission context on every user messageAll 8 hooks across 7 event types registered in hooks/hooks.json via direct python3 invocation. Single-process execution per hook (no subprocess chains).
| Defense | What it blocks |
|---|---|
| Phase Lock | Tool calls that don't match the current phase |
| Completion Guard | Completing without a validator report (relentless: requires Verdict: PASS) |
| Cleanup Guard | Deactivating without summary.md + clean worker-logs |
| Worker Test Block | Workers writing *.test.*, *.spec.*, *_test.*, tests/*, __tests__/* |
| Validator Path Lock | Validators writing to .mission/ except .mission/reports/* |
| Anti-Premature Stop | Completion in relentless mode when report says FAIL |
| Model Enforcement | Wrong model on Agent dispatch (auto-injects if missing) |
| Unknown Phase Block | All tool calls when phase is not in valid set |
| Symlink Protection | Path traversal via symlinks (realpath resolution) |
| Phase Transitions | Invalid transitions (e.g., worker → complete) |
| Stop Guard | Droid stopping while mission active (non-complete phase) |
| SubagentStop Guard | Workers/Validators stopping without handoff/report |
The Orchestrator creates .mission/features.json as the structured tracking system:
npx claudepluginhub ryan-mt/missionMulti-agent orchestration system for Claude Code
Real-time statusline HUD for Claude Code - context health, tool activity, agent tracking, and todo progress
Autonomous Development Orchestrator - Transform ideas into production-ready applications through multi-agent pipeline. Spec → Plan → Tasks → 100% Working App.
This skill should be used when the model's ROLE_TYPE is orchestrator and needs to delegate tasks to specialist sub-agents. Provides scientific delegation framework ensuring world-building context (WHERE, WHAT, WHY) while preserving agent autonomy in implementation decisions (HOW). Use when planning task delegation, structuring sub-agent prompts, or coordinating multi-agent workflows.
Multi-agent orchestration system for Claude Code
Conductor v3 — Multi-agent orchestration with Evaluate-Loop, parallel execution, Board of Directors, and bundled SupaConductor skills for Claude Code
Describe your goal, approve the spec, then step away — Claude and Codex loop together until it's right.
Production-ready Claude Code configuration with role-based workflows (PM→Lead→Designer→Dev→QA), safety hooks, 44 commands, 19 skills, 8 agents, 43 rules, 30 hook scripts across 19 events, auto-learning pipeline, hook profiles, and multi-language coding standards