From productionos
ProductionOS smart router — single entry point that routes to the right pipeline based on intent. The ONLY command new users need to know.
npx claudepluginhub shaheerkhawaja/productionos --plugin productionosThis skill uses the workspace's default tool permissions.
ProductionOS smart router — single entry point that routes to the right pipeline based on intent. The ONLY command new users need to know.
Searches, retrieves, and installs Agent Skills from prompts.chat registry using MCP tools like search_skills and get_skill. Activates for finding skills, browsing catalogs, or extending Claude.
Searches prompts.chat for AI prompt templates by keyword or category, retrieves by ID with variable handling, and improves prompts via AI. Use for discovering or enhancing prompts.
Guides agent creation for Claude Code plugins with file templates, frontmatter specs (name, description, model), triggering examples, system prompts, and best practices.
ProductionOS smart router — single entry point that routes to the right pipeline based on intent. The ONLY command new users need to know.
| Parameter | Values | Default | Description |
|---|---|---|---|
intent | string | required | What you want to do. Natural language or keyword. Examples: 'audit this project', 'fix the frontend', 'research authentication', 'review my PR', 'ship it', 'debug the login bug' |
target | string | -- | Target directory, file, or URL (default: current directory) |
You are the ProductionOS entry point. You receive a natural language intent and route to the right command.
This is the ONLY command a new user needs to know. Everything else is discoverable through this.
Run the shared ProductionOS preamble (templates/PREAMBLE.md).
Before static intent classification, run the agent index for intelligent agent selection:
bun run "${CLAUDE_PLUGIN_ROOT}/scripts/agent-index.ts" --goal "$ARGUMENTS.intent" 2>/dev/null
Parse the JSON output. If lowConfidence is false, the router has high-confidence agent matches:
ProductionOS Router → ${agents.length} agents matched (confidence: ${agents[0].confidence})If lowConfidence is true or the script fails, fall through to static keyword matching below.
Also detect the project stack for tool provisioning:
bun run "${CLAUDE_PLUGIN_ROOT}/scripts/stack-detector.ts" 2>/dev/null
Use the detected stack to inform tool selection in the routed command.
Classify $ARGUMENTS.intent into one of these categories:
Triggers: "audit", "review", "check", "scan", "grade", "score", "how good is"
Route to: /production-upgrade $ARGUMENTS.target
Triggers: "design", "redesign", "mockup", "UI", "UX", "frontend", "look", "feel", "ugly", "beautiful"
Route to: /designer-upgrade $ARGUMENTS.target
Triggers: "user story", "journey", "friction", "experience", "onboarding", "flow", "persona"
Route to: /ux-genie $ARGUMENTS.target
Triggers: "fix", "improve", "upgrade", "make better", "optimize", "refactor"
Route to: /auto-swarm-nth "$ARGUMENTS.intent" --target $ARGUMENTS.target
Triggers: "plan", "architect", "design system", "strategy", "roadmap", "vision"
Route to: /omni-plan-nth $ARGUMENTS.target
Triggers: "build", "create", "add", "implement", "feature", "new"
Route to: /brainstorming $ARGUMENTS.intent → then /writing-plans → then /auto-swarm-nth
Triggers: "debug", "bug", "broken", "error", "crash", "failing", "not working"
Route to: /debug $ARGUMENTS.intent
Triggers: "test", "QA", "verify", "check", "regression", "e2e"
Route to: /qa $ARGUMENTS.target
Triggers: "review PR", "review code", "review changes", "review diff"
Route to: /review
Triggers: "ship", "deploy", "release", "merge", "push", "PR"
Route to: /ship
Triggers: "research", "investigate", "explore", "find out", "compare", "what is"
Route to: /deep-research $ARGUMENTS.intent
Triggers: "evaluate", "self-eval", "how did I do", "score my work", "rate"
Route to: /self-eval session
Triggers: "help", "how do I", "what can you do", "list commands", "getting started"
Route to: /productionos-help
Before executing, confirm:
ProductionOS → Detected intent: {CATEGORY}
Routing to: /{command} {arguments}
Target: {target or "current directory"}
If the classification is ambiguous, ask the user to clarify:
I'm not sure if you want:
A) /production-upgrade (audit and fix code quality)
B) /designer-upgrade (audit and fix design/UI)
C) /auto-swarm-nth (parallel targeted fixes)
Which fits better?
Invoke the selected command via the Skill tool or by executing its protocol directly.
After the routed command completes, run self-eval:
/self-eval last
/build-productionos "audit this project" → Full codebase audit (/production-upgrade)
/build-productionos "fix the frontend" → UI/UX redesign (/designer-upgrade)
/build-productionos "research auth patterns" → Deep research (/deep-research)
/build-productionos "review my PR" → Code review (/review)
/build-productionos "ship it" → Ship workflow (/ship)
/build-productionos "debug login bug" → Systematic debugging (/debug)
/build-productionos "plan the new feature" → Strategic planning (/omni-plan-nth)
/build-productionos "test everything" → QA testing (/qa)
/build-productionos "build user profiles" → Feature development (brainstorm → plan → build)
/build-productionos "how did I do" → Self-evaluation (/self-eval session)
/build-productionos "help" → Usage guide (/productionos-help)
| Scenario | Action |
|---|---|
| No target provided | Ask for clarification with examples |
| Target not found | Search for alternatives, suggest closest match |
| Agent dispatch fails | Fall back to manual execution, report the error |
| Ambiguous input | Present options, ask user to pick |
| Execution timeout | Save partial results, report what completed |