skill-factory-router

Skill Factory Router

Select the smallest durable Skill Factory workflow for a Codex skill-management request, then return a read-only routing handoff before loading deeper instructions.

When To Use

• The user explicitly asks which lane owns a Codex skill-management request.
• The request names skill work but the correct Skill Factory lane is not already certain.
• The request depends on Tessl, Plugin Eval, validation, runtime, installation, package visibility, or session evidence and needs the correct owner lane before action.

Do not use this router to execute the downstream lane. Route plugin package lifecycle work to plugin-factory. Route system creator and installer work to the system lanes instead of copying those bodies into Skill Factory.

Example requests that should trigger this router:

• "Create a skill from this workflow and tell me which lane owns it."
• "This skill is failing package proof; route the smallest hardening workflow."
• "Should this be skill-refactor, skillify, or the system skill creator?"

Start with the smallest package boundary that can answer the routing question. Inspect the target request, target skill, and one evidence handle before expanding to broader package or runtime state.

Inputs

• User request or named target skill.
• Current authority boundary: read-only routing, approved edit, install or sync, eval execution, package publication, or review-only.
• Available evidence handle when routing depends on Tessl, Plugin Eval, validation, runtime, review, or session artifacts.

Ask one plain-language question only when missing information prevents lane selection. Read-only routing does not require edit authority or a validation command. Load references/discovery-interview.md only for unresolved routing ambiguity; do not turn a clear downstream request into an interview.

Outputs

Return one YAML handoff:

schema_version: 1
selected_lane: .system/skill-creator|skillify|skill-builder|skill-refactor|.system/skill-installer
mode: create|capture|harden|analyze|install
rationale: <one sentence tied to the request shape>
next_step: <specific skill or system lane to load next>
first_principles_check:
  required: true|false
  result: skill|docs|script|hook|validator|rule|answer|not_checked
blocked_by: null