cuioss/task-verification

Task Verification Skill

Role: Task executor skill for verification-only tasks (profile=verification). Loaded by plan-marshall:phase-5-execute skill when task.profile is verification.

Key Pattern: No files are modified. Steps contain verification commands to run. Each step is executed and marked done/failed. No domain skills are needed.

Contract Compliance

MANDATORY: Follow the execution contract defined in:

| Contract | Location | Purpose | |----------|----------|---------| | Task Contract | plan-marshall:manage-tasks/standards/task-contract.md | Task structure and fields |

Input

| Parameter | Type | Required | Description | |-----------|------|----------|-------------| | plan_id | string | Yes | Plan identifier | | task_number | number | Yes | Task number to execute |

Output

status: success | error
plan_id: {echo}
task_number: {echo}
execution_summary:
  steps_completed: N
  steps_total: M
  commands_run: [commands]
verification:
  passed: true | false
  command: "{cmd}"
next_action: task_complete | requires_attention
message: {error message if status=error}

Workflow

Step 1: Load Task Context

Read the task file:

python3 .plan/execute-script.py plan-marshall:manage-tasks:manage-tasks get \
  --plan-id {plan_id} \
  --number {task_number}

Extract key fields:

• profile: Should be verification
• steps: Verification commands to run
• verification: Verification criteria

Step 2: Execute Verification Steps

For each step (verification command):

1. Run the command:

{step.target}

2. Check exit code and output
3. Mark step complete:

python3 .plan/execute-script.py plan-marshall:manage-tasks:manage-tasks finalize-step \
  --plan-id {plan_id} \
  --task {task_number} \
  --step {N} \
  --outcome done

Step 3: Handle Failures

If a command fails:

1. Analyze error output
2. This is a verification task — do NOT modify source files
3. Report the failure with structured output for triage

If verification fails:

python3 .plan/execute-script.py plan-marshall:manage-tasks:manage-tasks update \
  --plan-id {plan_id} \
  --number {task_number} \
  --status blocked

Step 4: Return Results

On success:

status: success
plan_id: {plan_id}
task_number: {task_number}
execution_summary:
  steps_completed: {N}
  steps_total: {M}
  commands_run:
    - {cmd1}
verification:
  passed: true
  command: "{verification command}"
next_action: task_complete

On failure (structured output for phase-5-execute triage):

status: success
plan_id: {plan_id}
task_number: {task_number}
execution_summary:
  steps_completed: {N}
  steps_total: {M}
  commands_run:
    - {cmd1}
verification:
  passed: false
  command: "{failed command}"
  exit_code: {exit_code}
  stderr: "{truncated stderr, max 2000 chars}"
  findings:
    - type: {compile-error|test-failure|lint-issue}
      file: {file_path}
      line: {line_number}
      message: "{error message}"
next_action: requires_triage

The findings array is best-effort: parse compiler errors, test failures, or lint output into structured entries. If parsing fails, include the raw stderr for the triage step to analyze.

Integration

Invoked by: plan-marshall:phase-5-execute skill (when task.profile = verification)

Skill Loading: Agent resolves this skill via resolve-task-executor --profile verification

Script Notations (use EXACTLY as shown):

• plan-marshall:manage-tasks:manage-tasks - Task operations (get, update, finalize-step)