Validate and score AI agent skill files (SKILL.md, AGENTS.md) before publishing to the SkillFlow marketplace.
- Structure checks — Required sections, header hierarchy, content length
- Quality scoring — 0-100 score with letter grade (A-F)
- Security scanning — Detects hardcoded secrets, dangerous commands
- Platform detection — Verifies platform compatibility declarations
- Best practices — License, contributing, contact information
# Lint a single file
python skillflow_linter.py SKILL.md
# Lint all skills in a directory
python skillflow_linter.py --dir ./skills/
# JSON output for CI/CD
python skillflow_linter.py SKILL.md --json| Rule | Severity | Description |
|---|---|---|
| SF001 | Error | Missing title (H1 header) |
| SF002 | Warning | No description section |
| SF003 | Warning | No usage/instructions section |
| SF004 | Warning | No examples or code blocks |
| SQ001 | Error/Warning | Content too short |
| SQ002 | Warning | Skipped header level |
| SP001 | Warning | No platform compatibility mentioned |
| SS001 | Error | Potential secret detected |
| SS002 | Warning | Dangerous command found |
# GitHub Actions
- name: Lint Skills
run: python skillflow_linter.py SKILL.md --jsonMIT