菜单
Enforces evidence-based completion claims. Use before claiming work is done, tests pass, or bugs are fixed. Requires running verification commands and confirming output before any success claims.
默认使用会先检查来源的 Prompt；你也可以切换为直接命令，或下载本地副本。
决定是否安装前，请先阅读 SKILL.md，以及 SkillsMP 当前展示的配套文件。
用 Codex 或 Claude 帮你安装 复制这段 Prompt，粘贴到 Codex、Claude 或其他助手里，让它检查 Skill 页面并帮你完成安装。
直接命令不会经过审查 Prompt；运行前请先检查来源。
npx skills add https://github.com/Delphine-L/claude_global --skill verification-before-completion命令会保持在同一行。复制前请横向滚动并检查完整内容。
想先保存到本地？可下载 SkillsMP 当前能够提供的文件。
Run shell commands bare — no decorative echo headers ("=== X ==="), no echo-then-cmd chains, no trailing "echo done". Use the Bash tool's description field for any narration. Triggers any time you're about to issue a Bash command.
Expert guide for managing Claude Code global skills and commands. Use when creating new skills, symlinking to projects, updating existing skills, or organizing the centralized skill repository.
Token optimization best practices for cost-effective Claude Code usage. Automatically applies efficient file reading, command execution, and output handling strategies. Includes model selection guidance (Opus for learning, Sonnet for development/debugging). Prefers bash commands over reading files.
基于 SOC 职业分类
正在显示 SKILL.md
| name | verification-before-completion |
| description | Enforces evidence-based completion claims. Use before claiming work is done, tests pass, or bugs are fixed. Requires running verification commands and confirming output before any success claims. |
| version | 1.0.0 |
| allowed-tools | Read, Grep, Glob, Bash |
Core principle: Evidence before claims, always.
BEFORE claiming any status:
1. IDENTIFY: What command proves this claim?
2. RUN: Execute the command (fresh, complete)
3. READ: Full output, check exit code
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
| Claim | Requires | Not Sufficient |
|---|---|---|
| Tests pass | Test command output: 0 failures | Previous run, "should pass" |
| Linter clean | Linter output: 0 errors | Partial check |
| Build succeeds | Build command: exit 0 | "Looks good" |
| Bug fixed | Original symptom: gone | Code changed, assumed fixed |
| Pipeline complete | Output files exist and are valid | Tool reported success |
| Notebook runs | All cells executed without error | Some cells ran |
If you catch yourself:
| Excuse | Reality |
|---|---|
| "Should work now" | RUN the verification |
| "I'm confident" | Confidence != evidence |
| "Just this once" | No exceptions |
| "Tool said success" | Verify independently |
| "Partial check is enough" | Partial proves nothing |
Tests:
CORRECT: [Run test] → [See: 34/34 pass] → "All tests pass"
WRONG: "Should pass now" / "Looks correct"
Pipeline output:
CORRECT: [Check output file exists] → [Verify size/format] → "Pipeline produced valid output"
WRONG: "Galaxy shows green" (without checking actual data)
Notebook:
CORRECT: [Restart kernel] → [Run All] → [Check no errors] → "Notebook runs cleanly"
WRONG: "I updated the cell" (without re-running)
Run the command. Read the output. THEN claim the result.
Adapted from obra/superpowers verification-before-completion skill.