Automated Assessment Skill

Systematic compliance assessment for projects against checkpoint-enabled skills.

Running the Assessment

/assess                              # Assess against all matching skills
/assess skill-repo typo3-testing     # Assess against specific skills only
/assess --force                      # Run all skills, ignore preconditions
/assess --mechanical-only            # Skip LLM reviews, only scripted checks
/assess --autofix                    # Find and fix issues automatically
/assess skill-repo --autofix        # Fix specific skill's issues only
/assess --review                     # Categorize failures as skill improvements
/assess --autoimprove                # Fix + propose skill improvements

Options

Flag	Effect
`<skill-names>`	Only run checkpoints for named skills
`--force`	Skip precondition checks, run all skills
`--mechanical-only`	Skip LLM reviews, only run scripted checks
`--autofix`	Fix failures by invoking the responsible skill, then re-verify
`--review`	Categorize failures and suggest skill improvements
`--autoimprove`	Fix what's possible, propose improvements for the rest
`--create-issues`	With --autoimprove, create GitHub issues in skill repos
`--json`	Output raw JSON instead of formatted report

Steps

Discover matching skills, evaluate preconditions
Run scripted checks (mechanical checkpoints)
Group LLM checkpoints by domain, spawn parallel agents
Collect results and generate compliance report

Checkpoint Types

Mechanical: file_exists, file_not_exists, contains, not_contains, regex, json_path, gh_api, command. LLM: llm_review (grouped by domain). See references/checkpoints-schema.md.

Domain Groups

Domain	Focus
`repo-health`	README, badges, branding, AGENTS.md
`security`	SLSA, OpenSSF, SBOM, vulnerabilities
`code-quality`	PHPStan, tests, PHP 8.x patterns
`documentation`	RST, rendering, docs.typo3.org
`git-workflow`	Branching, commits, conventional commits
`docker`	Dockerfile, compose, container patterns
`ddev`	DDEV configuration, services, commands
`upgrade`	TYPO3 version upgrades, deprecations

Autofix Workflow

With --autofix: run checks, invoke responsible skill for failures (e.g., /agent-rules), re-verify. Statuses: auto-fixed (now passes), needs-review (LLM item), unfixable (still fails).

Review & Improvement

--review and --autoimprove create a feedback loop from assessment back into skills.

Category	Meaning	Action
`fixable`	A skill can fix this	`--autofix` runs the skill
`skill-gap`	Skill doesn't cover this	Propose SKILL.md update
`checkpoint-issue`	Checkpoint miscalibrated	Propose checkpoint change

--autoimprove runs autofix first, then for remaining failures: analyzes root cause, proposes changes (checkpoint severity/precondition/description), and outputs improvement JSON. Use --create-issues to file issues in skill repos.

Severity Levels

Level	Meaning	Action
`error`	Must fix before release	Blocks release
`warning`	Should fix	Recommendation
`info`	Nice to have	Optional

References

references/checkpoints-schema.md -- checkpoint YAML schema
references/checkpoint-workflow.md -- discovery, agents, validation
references/migration-guide.md -- adding checkpoints to skills

automated-assessment