agent-acceptance-gate - Multi-agent pre-merge acceptance gate
Verify multi-agent round outputs against success criteria, risks, and budget before merging
Tags
Updated: 2026-05-09Capabilities
Typical Inputs
Typical Outputs
What this skill does
- read plan file
- run verification commands
- check assertions
- read agent conversation
- check reconciler verdict
- aggregate risks
- audit prose changes
- check token budget
- write verdict report
Inputs
- .coord/plan.yml
- .ai/<agent>_log_<NNN>_<slug>.txt.result.json
- .coord/reconciliation_<NNN>.md
- changed prose files
Outputs
- .coord/acceptance_<NNN>.md
- per-task pass/fail status
- risk summary
- budget compliance status
- overall pass/fail/conditional verdict
Requirements
- .coord/plan.yml file
- multi-agent round completed
- reconciler output available
- optional: academic-writing-skills