Agent skill

task-verification

MANDATORY skill. Invoke BEFORE claiming any task complete. No exceptions. TRIGGERS: creating/modifying ANY file (code, script, config, docs, data). Simple changes like renaming variables, fixing typos, adding comments ALL require verification. CORE RULE: Verification = running a Bash command (python -m py_compile, node --check, etc). Reading file back is NOT verification. "Should work" is NOT verification. If you cannot show command output as evidence, the task is NOT complete. Use this skill whenever you write, edit, fix, or create anything.

View SKILL.md on GitHub Repository
Stars 163
Forks 31

Install this agent skill to your Project

npx add-skill https://github.com/majiayu000/claude-skill-registry/tree/main/skills/other/other/task-verification

SKILL.md

Task Verification - MANDATORY

STOP! READ THIS BEFORE CLAIMING COMPLETE

┌─────────────────────────────────────────────────────────────────────┐
│                                                                     │
│   IF YOU DID NOT RUN A BASH/SHELL COMMAND TO VERIFY,               │
│   YOU ARE NOT DONE. STOP HERE AND VERIFY NOW.                      │
│                                                                     │
│   Reading a file with the Read tool ≠ verification                 │
│   Using cat to show file contents ≠ verification                   │
│   Thinking "it looks correct" ≠ verification                       │
│                                                                     │
│   VERIFICATION = RUNNING BASH COMMAND + CAPTURING OUTPUT           │
│                                                                     │
└─────────────────────────────────────────────────────────────────────┘

The Verification Requirement

What You MUST Do

  1. Create/modify your file
  2. RUN A BASH COMMAND (not Read tool, not cat, but an actual verification command)
  3. Capture the bash output
  4. Include the output in your response

Minimum Verification Commands by File Type

File Type REQUIRED Bash Command
Any .py file python -m py_compile file.py
Executable .py script python file.py [test_args]
.js file node --check file.js
.json file python -c "import json; json.load(open('file.json')); print('✓')"
.yaml/.yml python -c "import yaml; yaml.safe_load(open('file.yaml')); print('✓')"
.sh file bash -n file.sh

Examples of CORRECT Verification

Example: Variable Rename

Task: Change x = 0 to count = 0

YOUR RESPONSE MUST INCLUDE:

Created code.py with count = 0

Verification:
$ python -m py_compile code.py && echo "Syntax OK"
Syntax OK

✓ Verified

WRONG - Do NOT do this:

Created code.py with count = 0
Verification: read the file back, it shows count = 0

↑ This is NOT verification. You must RUN a command.

Example: New Script

Task: Create factorial script

YOUR RESPONSE MUST INCLUDE:

Created factorial.py

Verification:
$ python factorial.py 5
120

$ python factorial.py 0
1

✓ Verified

The Bash Command Requirement

CRITICAL: Verification MUST use the Bash tool to run a command.

NOT acceptable:

  • Read tool to read file back
  • cat file.py to show contents
  • head file.py or tail file.py
  • Saying "I verified by reading"

REQUIRED:

  • python -m py_compile file.py
  • python file.py args
  • python -c "import json; json.load(...)"
  • node file.js
  • pytest tests/

Completion Checklist

Before saying "complete", verify ALL:

  • I used the Bash tool to run a verification command
  • The command I ran was NOT cat or head or tail
  • I captured the actual output from bash
  • I included the bash output in my response
  • The output proves my work is correct

Decision Flow

You made a change to a file.
        │
        ▼
┌────────────────────────────────────────┐
│ Did you use Bash tool to run          │
│ python -m py_compile or similar?      │
└────────────────────────────────────────┘
        │
        ├── NO → STOP! You are NOT done.
        │          Run: python -m py_compile your_file.py
        │
        └── YES → Include the output. NOW you can be done.

Quick Commands to Copy

bash
# Python syntax check - use for ANY .py file
python -m py_compile your_file.py && echo "✓ Syntax OK"

# Run Python script
python your_script.py

# JSON validation
python -c "import json; json.load(open('file.json')); print('✓ Valid JSON')"

# YAML validation
python -c "import yaml; yaml.safe_load(open('file.yaml')); print('✓ Valid YAML')"

FINAL REMINDER: If your "verification" involved the Read tool or cat command, you have NOT verified. Run an actual command with Bash.

Recommended Agent Skills

Expand your agent's capabilities with these related and highly-rated skills.

majiayu000/claude-skill-registry

agent-ops-spec

Manage specification documents in .agent/specs/. Use when user provides requirements, acceptance criteria, or feature descriptions that need to be tracked and validated against implementation.

163 31
Explore
majiayu000/claude-skill-registry

agent-ops-state

Maintain .agent state files. Use at session start, after meaningful steps, and before concluding: read/update constitution/memory/focus/issues/baseline consistently.

163 31
Explore
majiayu000/claude-skill-registry

agent-ops-spec

Manage specification documents in .agent/specs/. Use when user provides requirements, acceptance criteria, or feature descriptions that need to be tracked and validated against implementation.

163 31
Explore
majiayu000/claude-skill-registry

agent-ops-testing

Test strategy, execution, and coverage analysis. Use when designing tests, running test suites, or analyzing test results beyond baseline checks.

163 31
Explore
majiayu000/claude-skill-registry

agent-ops-testing

Test strategy, execution, and coverage analysis. Use when designing tests, running test suites, or analyzing test results beyond baseline checks.

163 31
Explore
majiayu000/claude-skill-registry

agent-ops-state

Maintain .agent state files. Use at session start, after meaningful steps, and before concluding: read/update constitution/memory/focus/issues/baseline consistently.

163 31
Explore

Didn't find tool you were looking for?

Be as detailed as possible for better results