plan-secure

原文涵盖的主题

• plan-secure workflow
• Overview
• Invocation
• How to invoke

Run harnessed prompt plan-secure --task "$ARGUMENTS" --json, execute /gsd-secure-phase, then complete the registered checkpoint with verifiable output. Fires only when phase.hasauthorsecrets == true. !harnessed checkpoint intent plan-secure The banner above (when present) means this invocation is REGISTERED with the engine (an intent marker) — not yet compliant: the steps below (prompt → spawn → checkpoint complete) resolve it, and a per-turn reminder persists until they run. The numbered sequence below is the state machine — execute it with Bash. Do NOT improvise an equivalent flow from the Overview above: freelancing bypasses the engine (no ledger, no evidence guard). harnessed gives you the spawn-ready prompt; YOU spawn the subagent with a {{ host.native }} {{ host.spawnsubagent }} (keeps the session responsive + lets clarification round-trips reach the user). Do NOT pipe to harnessed run plan-secure — that is the CI/headless path ({{ host.harnessedrunwarningnote.executiontail }} 1. Bash: harnessed prompt plan-secure --task "$ARGUMENTS" --json → parse {prompt, maxiterations, model, resultfile}. 2. Spawn a {{ host.native }} subagent ({{ host.spawnsubagent }}) with that prompt and

这里只展示节选，请以完整原文和当前要求为准。