sloth-verify

name	sloth-verify
description	Use when about to claim a Maycrest deliverable is complete, fixed, or passing — before committing, opening a PR, or telling the client, requiring fresh verification evidence first.

Sloth Verify — Evidence Before Claims 🦥

Overview

Telling a client work is done without verifying it is dishonesty, not efficiency. At Maycrest the deliverable is trust as much as code — a false "it works" costs us the relationship.

Core principle: Evidence before claims, always.

Violating the letter of this rule is violating the spirit of this rule.

The Iron Law

NO COMPLETION CLAIMS WITHOUT FRESH VERIFICATION EVIDENCE

If you haven't run the verification command in this message, you cannot claim it passes.

The Gate Function

BEFORE claiming any status or expressing satisfaction:

1. IDENTIFY: What command proves this claim?
2. RUN: Execute the FULL command (fresh, complete)
3. READ: Full output, check exit code, count failures
4. VERIFY: Does output confirm the claim?
   - If NO: State actual status with evidence
   - If YES: State claim WITH evidence
5. ONLY THEN: Make the claim

Skip any step = lying, not verifying

Common Failures

Claim	Requires	Not Sufficient
Tests pass	Test command output: 0 failures	Previous run, "should pass"
Linter clean	Linter output: 0 errors	Partial check, extrapolation
Build succeeds	Build command: exit 0	Linter passing, logs look good
tsc clean	tsc --noEmit: 0 errors	Editor squiggles gone
Bug fixed	Test original symptom: passes	Code changed, assumed fixed
Regression test works	Red-green cycle verified	Test passes once
Subagent completed	Git diff shows the changes	Agent reports "success"
Requirements met	Line-by-line checklist vs. spec	Tests passing

Forbidden Phrases — STOP

Never ship these unless you've just run the verification:

• "should work now", "probably", "seems to"
• Expressing satisfaction before evidence ("Great!", "Perfect!", "Done!")
• About to commit/push/PR without verification
• Trusting a subagent's success report
• Relying on partial verification
• "just this once" / "I'm tired, ship it"
• ANY wording implying success without having run verification

Rationalization Prevention

Excuse	Reality
"Should work now"	RUN the verification
"I'm confident"	Confidence ≠ evidence
"Just this once"	No exceptions
"Linter passed"	Linter ≠ compiler
"The client's waiting"	A false claim costs more than 30 seconds
"Subagent said success"	Verify independently via the diff
"I'm tired"	Exhaustion ≠ excuse
"Different words so the rule doesn't apply"	Spirit over letter

Key Patterns

Tests:

✅ [Run test command] [See: 165/165 pass] "All tests pass"
❌ "Should pass now" / "Looks correct"

Regression tests (TDD red-green):

✅ Write → Run (pass) → Revert fix → Run (MUST FAIL) → Restore → Run (pass)
❌ "I've written a regression test" (without red-green verification)

Build:

✅ [Run npm run build] [See: exit 0] "Build passes"
❌ "Linter passed" (linter doesn't check compilation)

Requirements:

✅ Re-read spec → Build checklist → Verify each → Report gaps or completion
❌ "Tests pass, phase complete"

Subagent delegation:

✅ Agent reports success → Check git diff → Verify changes exist → Report actual state
❌ Trust the agent's report

Why This Matters

• Corey says "I don't believe you" — trust broken, and that's the whole business.
• Undefined functions shipped to a client — the app crashes on their device.
• Missing requirements shipped — incomplete deliverable, awkward redo on the call.
• Time wasted on a false completion → client redirect → rework on our dime.

When To Apply

ALWAYS before:

• ANY variation of a success/completion claim
• ANY expression of satisfaction
• Committing, opening a PR, marking a task done
• Moving to the next task
• Handing a build to the client or to maycrest-ops:reality-checker

Rule applies to: exact phrases, paraphrases, synonyms, implications of success — ANY communication suggesting completion or correctness.

Example: AOS Sober Living booking app

About to tell Corey the van-booking fix is done. The gate:

1. Identify — npx vitest run src/booking.test.ts proves the double-booking guard.
2. Run — fresh, full command.
3. Read — Test Files 1 passed, Tests 4 passed (4), exit 0.
4. Verify — output confirms the claim, including the original-symptom test.
5. Claim — "Double-booking is fixed: 4/4 booking tests pass, build exits 0." Evidence attached.

Without step 2-3, the only honest statement is "I changed the code but haven't verified it."

Integration with the Maycrest roster

• Runs on every task inside sloth-build and after sloth-tdd green.
• Re-check the plan from sloth-blueprint line-by-line for requirements coverage.
• For the production-readiness verdict, hand the verified build to maycrest-ops:reality-checker — it defaults to NEEDS WORK and demands the same evidence.

The Bottom Line

No shortcuts for verification. Run the command. Read the output. THEN claim the result. This is non-negotiable.

---

Adapted from the MIT-licensed obra/superpowers project (© 2025 Jesse Vincent). See NOTICE.