Stage 6: Documentation & SKILL.md#

Objective#

Run tests, update TEST.md with results, and generate SKILL.md for agent discovery.

Inputs#

• Test code from Stage 5
• Source code from Stage 3

Steps#

1. Run tests:
   • Python: pytest tests/ -v --tb=short
   • Node.js: node --test tests/ or npx jest tests/
   • Record pass/fail counts and any failures

1. Update TEST.md (Part 2) — append test results:

   ## Test Results
   - Total: N tests
   - Passed: N
   - Failed: N
   - Skipped: N

   ### Failures
   - test_xxx: <failure reason>

1. Generate SKILL.md:

   ---
   name: cli-anything-<SOFTWARE>
   description: <one-line description for LLM tool selection>
   version: 1.0.0
   tags: [<SOFTWARE>, cli-anything, agent-tool]
   entry_point: cli-anything-<SOFTWARE>
   ---

   # <SOFTWARE> Agent Tool

   You have access to `cli-anything-<SOFTWARE>`, a command-line tool for AI agent control of <SOFTWARE>.

   ## Available Commands

   | Command | Description | Example |
   |---------|-------------|---------|
   | `project create` | Create new project | `cli-anything-<SOFTWARE> project create --name myproject` |
   | ... | ... | ... |

   ## Usage Patterns

   ### Basic Workflow
   ```bash
   cli-anything-<SOFTWARE> project create --name demo --json
   cli-anything-<SOFTWARE> <domain> <action> --json
   cli-anything-<SOFTWARE> export render --output result.png --json

## Error Handling

If a command fails, check --json output for {"status": "error", "error": "..."}.

## Output
- Updated `TEST.md` (Part 2)
- `skills/SKILL.md`