Update to RUMBO Agent Reliability

{
  "name": "execute-verify-close",
  "description": "Execute an authorized bounded action, verify the actual effect, and close only with evidence.",
  "included_files": [],
  "skill_md_contents": "---\nname: execute-verify-close\ndescription: Execute an authorized bounded action, verify the actual effect, and close only with evidence.\n---\n\n# Execute Verify Close\n\nUse only when exact target, intended effect, authorization, rollback/no-effect path, and verification criterion are known.\n\nState machine:\nREADY -> AUTHORIZED -> EXECUTED -> VERIFIED -> CLOSED\n\nFailure path:\nEXECUTED -> FAILED -> RECONCILE\n\nNever infer success from process exit, queue acceptance, UI feedback, model output, or a prepared artifact.\n\nVerify:\n- exact target\n- expected postcondition\n- unintended changes\n- execution/effect receipt\n- final authoritative state\n- rollback/no-effect status\n\nUse NOT_PROVEN when the effect cannot be independently evidenced. Only mark CLOSED when required evidence exists.\n"
}