메뉴
Route tasks to specific models using the repository routing.json policy and log each decision to routing_ledger.md. Use when the user asks to dispatch, orchestrate, route, or when multi-model routing is required.
Codex 또는 Claude로 설치 이 Prompt를 복사해 Codex, Claude 또는 다른 어시스턴트에 붙여 넣으면 Skill 페이지를 검토하고 설치를 진행할 수 있습니다.
SOC 직업 분류 기준
Match emotional frequency, detect power dynamics (444_CONSULT). Detects emotional state and power imbalance. The empathy detection stage of arifOS metabolic loop. Use when emotional calibration matters.
Establish position, intake context, ground reality (000_INTAKE). The intake stage of arifOS metabolic loop. Reduces sensory entropy by fixing a reference frame. Use at the beginning of any interaction to establish grounding.
Check F1-F9 floors, compute governance score (777_VERIFY). Checks constitutional floors. The enforcement layer of APEX tier. Use to verify any action against arifOS constitution.
Reduce entropy, refine output, cool structure (888_HOLD). Reduces entropy and refines output. The cooling process. Use when refining drafts or reducing complexity.
Combine knowledge, cross-domain synthesis (333_DELIBERATE). Combines knowledge across domains and forges creative connections. Use when connecting disparate domains or synthesizing insights.
Logical inference, causal trace, plan steps (222_PROCESS). Executes logical inference and causal analysis. The cognitive engine of AGI tier. Use when breaking down complex problems into logical steps.
| name | route-tasks-by-policy |
| description | Route tasks to specific models using the repository routing.json policy and log each decision to routing_ledger.md. Use when the user asks to dispatch, orchestrate, route, or when multi-model routing is required. |
routing.json from the repository root.keywords (case-insensitive) for each route.default_model and set task_type to unknown_default.routing_ledger.md before executing the task.Helper script:
scripts/route_task.py --prompt "<task>" [--log]
Use a single-line entry:
[timestamp] | TASK: <task_type> | MODEL: <model> | REASON: <reason>
routing.json unless the user explicitly instructs it.routing.json or routing_ledger.md unless asked.