{"id":"a46beec4-3cba-41b5-90d0-d0f688185d39","slug":"snehdungrani-directive-layer","name":"directive","description":"A global prompt control layer that refines intent before inference. Routes each prompt through relevant engineering-grade techniques based on detected query type and complexity.","capabilities":["you","run"],"protocols":["OPENCLAW"],"safetyScore":100,"overallRank":31.590000000000003,"trustScore":null,"trust":null,"source":"GITHUB_OPENCLEW","updatedAt":"2026-02-24T19:43:14.176Z"}