AgentSkillsCN

write-agent-report

为代理撰写报告时，务必选用恰当的命名与规范的存放位置。

SKILL.md
--- frontmatter
name: write-agent-report
description: Write an agent report with proper naming and location
license: MIT
compatibility: opencode
metadata:
  category: reporting

Write an agent report using the standard reporting script.

What This Command Does

Input: Report content and agent metadata

Output: A properly formatted agent report file

Creates report at: ~/.ai/wip/agent_reports/<agent_name>/<agent_id>-<date>.report.md

Workflow

Step 1: Gather Parameters

Required:

  • •agent_name: Name of the agent (e.g., "micro-tdd-agent", "task-extraction-agent")
  • •report_content: The markdown content of the report

Optional:

  • •agent_id: Unique identifier (default: auto-generated timestamp in YYYYMMDD_HHMMSS format)
  • •date: ISO date (default: today in YYYY-MM-DD format)

Step 2: Write Report

Use the write-agent-report.sh script via Bash tool:

bash
cat <<EOF | ~/.ai/scripts/generic/write-agent-report.sh <agent_name> [agent_id] [date]
<report_content>
EOF

The script will:

  • •Create the report directory if needed
  • •Generate agent_id and date if not provided
  • •Write the report with proper naming convention
  • •Return the full path to the created report

Step 3: Return Report Path

Return the full path to the created report:

code
Report written: <full_path_to_report>

Examples

Auto-generate agent_id and date

bash
cat <<EOF | ~/.ai/scripts/generic/write-agent-report.sh micro-tdd-agent
# Micro TDD Report - Test Name

## Status
✅ Success
EOF

Specify agent_id

bash
cat <<EOF | ~/.ai/scripts/generic/write-agent-report.sh micro-tdd-agent 20250129_143022
# Micro TDD Report - Test Name

## Status
✅ Success
EOF

Specify agent_id and date

bash
cat <<EOF | ~/.ai/scripts/generic/write-agent-report.sh micro-tdd-agent 20250129_143022 2025-01-29
# Micro TDD Report - Test Name

## Status
✅ Success
EOF

Report Content Guidelines

The report content should be markdown and typically includes:

For TDD Agents

  • •Test name and file
  • •Implementation details
  • •Verification steps
  • •Commit information
  • •Status (✅ Success or ❌ Failure)

For Extraction Agents

  • •Task information
  • •Framework phases completed
  • •Files created/modified
  • •Commit information
  • •Status

For Validation Agents

Use write-validation-report skill instead.

Success Criteria

  • •✅ Report written to correct location
  • •✅ Proper naming convention used
  • •✅ Full path returned
  • •✅ Report content is valid markdown