mnhat-executing

name	mnhat-executing
description	Implement one approved current-work item or parent-assigned slice. Use when planning or swarming has already chosen the scope and the next step is to make the change, verify it, and report precisely.

MNHAT Executing

Quick start

• Require exactly one approved current-work slice.
• Read the current planning and validation artifacts first.
• Make the change, verify it, and write .mnhat/<YYYY-MM-DD>-<feature-slug>/executing/execution-report.md.
• If a meaningful decision is made, call mnhat-implementation-record.

Workflows

1. Read AGENTS.md, context, plan, and validation report.
2. Confirm one approved current-work item and bounded scope.
3. Implement without drifting into adjacent work.
4. Run the required verification.
5. Write execution-report.md.
6. Pause only through .mnhat/state.json with explicit resume data.

Outputs

• Canonical artifact: .mnhat/<YYYY-MM-DD>-<feature-slug>/executing/execution-report.md
• Canonical template owner: references/execution-report-template.md

State rules

• execution-report.md is the primary artifact when state.json.phase is executing.
• artifact_paths[] should usually include the execution report, the implementation record when present, and the immediate input artifact such as current-story-pack.md or validation-report.md.
• If blocked by a failure that needs root-cause work, hand off to mnhat-debugging.

Further reading

• references/worker-details.md - result fields, pause rules, and scope checks
• references/execution-report-template.md - canonical execution-report.md form