Skip to content

/massu-loop

The implementation loop - execute plan items with mandatory verification until complete


/massu-loop

The /massu-loop command is the implementation engine. It takes a plan document and executes each deliverable with mandatory verification, looping until every item is complete. This is where code actually gets written, tested, and verified.

Usage

/massu-loop docs/plans/2026-02-13-user-profile.md

How It Works

The loop follows this cycle for each plan item:

1. Read the plan file (from disk, not memory)
2. Find the next incomplete deliverable
3. Implement the deliverable
4. Run the specified verification (VR-FILE, VR-GREP, VR-BUILD, VR-TEST)
5. Show verification command and output as PROOF
6. If verification fails: fix and re-verify
7. Mark item as complete
8. Go to step 1
9. When all items complete: run full verification suite

Mandatory Behaviors

Proof, Not Claims

For every completed item, the loop shows the verification command and its output:

## Deliverable D-001: ProfileForm.tsx

Created src/components/settings/ProfileForm.tsx

### Verification: VR-FILE
$ ls -la src/components/settings/ProfileForm.tsx
-rw-r--r-- 1 dev staff 2847 Feb 13 14:23 ProfileForm.tsx
VERIFIED

Negative Verification for Removals

If a plan item involves removing something, the loop runs negative verification:

## Deliverable D-008: Remove legacy profile component

Deleted src/components/legacy/OldProfile.tsx

### Verification: VR-NEGATIVE
$ grep -rn "OldProfile" src/
(no output - 0 matches)
VERIFIED: Pattern removed from codebase

All Items Required

The loop continues until every plan item is verified. Stopping at "most items" is a protocol violation. The loop must process every deliverable in the plan.

Re-Read Plan Each Iteration

The plan file is re-read from disk on each iteration, not from memory. This ensures any updates to the plan (from /massu-plan auditing) are picked up.

Final Verification

After all items are complete, the loop runs:

  1. npx tsc --noEmit -- Type safety check
  2. npm test -- All tests must pass
  3. npm run build -- Build must succeed
  4. Pattern scanner -- All patterns must comply

Tips

  • Long plans may require multiple sessions -- task linking ensures continuity
  • If a deliverable proves impossible, update the plan document first, then continue
  • The loop tracks progress in the memory database for cross-session visibility
  • Plan progress is visible via massu_memory_sessions