home / skills / hxk622 / tokendance / verification-before-completion

This skill verifies before completion by running the verification command and reporting evidence before claiming success.

npx playbooks add skill hxk622/tokendance --skill verification-before-completion

Review the files below or copy the command above to add this skill to your agents.

Files (1)
SKILL.md
4.4 KB
---
name: verification-before-completion
display_name: 完成前验证
version: 1.0.0
author: system
description: "在声称工作完成、修复或通过之前使用,在提交或创建 PR 之前。需要运行验证命令并确认输出后才能声称成功。证据先于断言。Use when about to claim work is complete - evidence before assertions always."
tags: [verification, verify, complete, 验证, 完成, test, 测试, check, 检查, evidence, 证据]
allowed_tools: []
max_iterations: 15
timeout: 300
enabled: true
match_threshold: 0.65
priority: 8
---

# Verification Before Completion

## Overview

Claiming work is complete without verification is dishonesty, not efficiency.

**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 |
| Bug fixed | Test original symptom: passes | Code changed, assumed fixed |
| Regression test works | Red-green cycle verified | Test passes once |
| Agent completed | VCS diff shows changes | Agent reports "success" |
| Requirements met | Line-by-line checklist | Tests passing |

## Red Flags - STOP

- Using "should", "probably", "seems to"
- Expressing satisfaction before verification ("Great!", "Perfect!", "Done!", etc.)
- About to commit/push/PR without verification
- Trusting agent success reports
- Relying on partial verification
- Thinking "just this once"
- Tired and wanting work over
- **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 |
| "Agent said success" | Verify independently |
| "I'm tired" | Exhaustion ≠ excuse |
| "Partial check is enough" | Partial proves nothing |
| "Different words so rule doesn't apply" | Spirit over letter |

## Key Patterns

**Tests:**
```
✅ [Run test command] [See: 34/34 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 build] [See: exit 0] "Build passes"
❌ "Linter passed" (linter doesn't check compilation)
```

**Requirements:**
```
✅ Re-read plan → Create checklist → Verify each → Report gaps or completion
❌ "Tests pass, phase complete"
```

**Agent delegation:**
```
✅ Agent reports success → Check VCS diff → Verify changes → Report actual state
❌ Trust agent report
```

## Why This Matters

From 24 failure memories:
- your human partner said "I don't believe you" - trust broken
- Undefined functions shipped - would crash
- Missing requirements shipped - incomplete features
- Time wasted on false completion → redirect → rework
- Violates: "Honesty is a core value. If you lie, you'll be replaced."

## When To Apply

**ALWAYS before:**
- ANY variation of success/completion claims
- ANY expression of satisfaction
- ANY positive statement about work state
- Committing, PR creation, task completion
- Moving to next task
- Delegating to agents

**Rule applies to:**
- Exact phrases
- Paraphrases and synonyms
- Implications of success
- ANY communication suggesting completion/correctness

## The Bottom Line

**No shortcuts for verification.**

Run the command. Read the output. THEN claim the result.

This is non-negotiable.

Overview

This skill enforces a strict verification-before-completion habit: never claim work is done until you run the definitive verification command and record its output. It prioritizes evidence over assertions so commits, PRs, and status updates are trustworthy. Follow the gate function: identify the proving command, run it fresh, inspect output and exit code, then report with evidence.

How this skill works

Before any claim of success, the skill guides you to identify the exact verification command that proves the claim, run it completely, and capture the full output and exit code. You then read and compare results against the expected success criteria (zero failures, exit 0, passing checks). Only after explicit verification do you make positive statements; otherwise you report the actual status with evidence.

When to use it

  • Before committing code or opening a pull request
  • Before declaring a bug fixed or a regression resolved
  • Before saying tests, build, or linter pass
  • Before moving to the next task or closing a ticket
  • Before delegating work to agents or relying on their reports

Best practices

  • Always run the full, intended verification command — not a subset or quick check
  • Capture and keep the raw output and exit code as evidence when reporting success
  • Perform red-green verification for regressions (force failure, restore fix, re-run tests)
  • Avoid subjective words like “should”, “probably”, or “seems” when status is unverified
  • Treat agent reports as leads — verify changes (VCS diff, tests) yourself before claiming success

Example use cases

  • Running the project test suite and pasting the test summary (e.g., 34/34 pass) before marking tests as passed
  • Executing the build command and including the exit 0 confirmation before merging
  • Verifying a bug fix by reproducing the original failing symptom, confirming failure, applying fix, and showing the passing run
  • Re-running a linter and showing zero errors before stating the codebase is lint-clean
  • Checking VCS diffs and running tests after an agent-made change before accepting the result

FAQ

What if verification is long-running or flaky?

Still run it. Capture timestamps, retry logs, and flaky test evidence. If flakiness prevents a definitive claim, report the instability with logs instead of asserting success.

Can I rely on a colleague or CI report?

Use colleague/CI reports as input but independently run or reproduce the verification when claiming completion. Attach CI logs as evidence only after you confirm them.