AgentSkillsCN

report

为已完成的任务模块生成结案报告。

SKILL.md
--- frontmatter
name: report
description: Generate a completion report for a finished task module
arguments:
  - name: task_module
    description: "Path to the task module directory (e.g., AiTasks/auth-refactor)"
    required: true
  - name: format
    description: "Report format: full (default) or summary"
    required: false
    default: full

/ai-cli-task report — Generate Completion Report

Generate a structured completion report for a task module, documenting what was planned, executed, and verified.

Usage

code
/ai-cli-task report <task_module_path> [--format full|summary]

Prerequisites

  • Task module should have status complete (post-exec assessment passed)
  • Can also be run on blocked or cancelled tasks for documentation purposes
  • Minimum content: If status is draft and no plan files exist, report outputs a brief notice ("No meaningful content to report — task is still in draft with no plan") instead of generating an empty report structure

Report Structure

Full Format

markdown
# Task Report: <title>

## Summary
- **Status**: complete | blocked | cancelled
- **Created**: <timestamp>
- **Completed**: <timestamp>
- **Duration**: <calculated>

## Objective
<!-- From .target.md -->

## Plan
<!-- Summary of implementation approach from plan files -->

## Changes Made
<!-- List of files modified/created/deleted with brief descriptions -->

## Verification
<!-- From .test/ criteria and results files, build status, evaluation outcomes -->

## Issues Encountered
<!-- From .bugfix/ if exists, or "None" -->

## Dependencies
<!-- Status of depends_on modules -->

## Lessons Learned
<!-- Any notable patterns, workarounds, or discoveries -->

Summary Format

Compact single-section report with: status, objective (1 line), key changes (bullet list), verification result.

Output

The report is written to AiTasks/<module_name>/.report.md and also printed to screen.

Execution Steps

  1. Read .index.md for task metadata (including completed_steps)
  2. Read .target.md for objectives
  3. Read all plan files for implementation approach
  4. Read .summary.md if exists (condensed context overview)
  5. Read .test/ for verification criteria and test results (all files, sorted by name, if exists)
  6. Read .analysis/ for evaluation history (all files, sorted by name, if exists)
  7. Read .bugfix/ for issue history (all files, sorted by name, if exists)
  8. Read .notes/ for research findings and experience log (all files, sorted by name, if exists)
  9. Collect git changes related to the task (if identifiable)
  10. Compose report in requested format
  11. Write to .report.md
  12. Distill experience: If task status is complete and type is non-empty, validate type matches [a-zA-Z0-9_:-]+ then extract key learnings and append to AiTasks/.experience/<type>.md (create file if not exists). Acquire AiTasks/.experience/.lock before writing (see Concurrency Protection in commands/ai-cli-task.md). Each entry is a section headed ### <module> (<date>) containing: what worked, what didn't, key decisions, tools/patterns discovered. If .experience/<type>.md exceeds 500 lines, compact it by summarizing older entries. Release lock after write
  13. Git commit: -- ai-cli-task(<module>):report generate completion report
  14. Write .auto-signal: { step: "report", result: "(generated)", next: "(stop)", checkpoint: "" }
  15. Print report to screen

Note: Report is a terminal step — it reads ALL history files (not just latest) to produce a comprehensive record. .summary.md is used as an overview, not a replacement for full history in report context.

State Transitions

No status change — report generation is informational. The task must already be complete, blocked, or cancelled.

Git

  • -- ai-cli-task(<module>):report generate completion report

.auto-signal

{ "step": "report", "result": "(generated)", "next": "(stop)", "checkpoint": "", "timestamp": "..." }

Report is always a terminal step — next is always (stop).

Notes

  • Reports are overwritten on regeneration (only latest report kept)
  • For blocked tasks, the report documents what was completed and what blocks remain
  • For cancelled tasks, the report documents the reason for cancellation
  • The report serves as a permanent record even after task files are archived
  • For complete tasks, report includes change history via git log --oneline --all --grep="ai-cli-task(<module>)" (uses commit message pattern, works even after task branch deletion)