mthds-check

name	mthds-check
description	Check and validate MTHDS bundles for issues. Use when user says "validate this", "check my workflow", "check my method", "does this .mthds make sense?", "review this pipeline", "any issues?", "is this correct?". Reports problems without modifying files. Read-only analysis.
min_mthds_version	0.17.0
allowed-tools	["Bash","Read","Grep","Glob"]

Check MTHDS bundles

Validate and review MTHDS bundles based on the MTHDS standard without making changes.

Process

Do not write .mthds files manually, do not do any other work. The CLI is required for validation, formatting, and execution — without it the output will be broken.

No backend setup needed: This skill works without configuring inference backends or API keys. You can start building/validating methods right away.

Step 1: Read the .mthds File

Load and parse the method.

Step 2: Run plxt Lint

Catch TOML syntax and schema errors before semantic validation. This skill is read-only and never triggers the PostToolUse hook, so lint must be run explicitly:

mthds-agent plxt lint <file>.mthds

If lint reports errors, include them in the final report and continue — semantic validation in the next step may reveal additional issues.

Step 3: Run CLI Validation

Use -L pointing to the bundle's own directory to avoid namespace collisions. --graph generates a flowchart:

mthds-agent validate bundle <file>.mthds -L <bundle-directory>/ --graph

Step 4: Parse the JSON Output

• If success: true — all pipes validated, report clean status
• If error — see Error Handling Reference for error types and recovery

Step 5: Cross-Domain Validation

When the bundle references pipes from other domains, use --library-dir (see Error Handling — Cross-Domain).

Step 6: Analyze for Additional Issues

Manual review beyond CLI validation:

• Unused concepts (defined but never referenced)
• Unreachable pipes (not in main_pipe execution path)
• Missing descriptions on pipes or concepts
• Inconsistent naming conventions
• Potential prompt issues (missing variables, unclear instructions)

Step 7: Report Findings

Report by severity:

• Errors: Validation failures from CLI (with error_type and pipe_code) and plxt lint errors
• Warnings: Issues that may cause problems (e.g., model availability)
• Suggestions: Improvements for maintainability
• Flowchart: If validation succeeded, mention the generated dry_run.html flowchart next to the bundle

Step 8: Do NOT Make Changes

This skill is read-only.

What Gets Checked

• TOML syntax and schema validation (via mthds-agent plxt lint)
• Concept definitions and references
• Pipe type configurations
• Input/output type matching
• Variable references in prompts
• Cross-domain references
• Naming convention compliance
• Model reference resolution (dry run)

Reference

• Error Handling — read when CLI returns an error to determine recovery
• MTHDS Agent Guide — read for CLI command syntax or output format details
• MTHDS Language Reference — read when reviewing .mthds TOML syntax
• Native Content Types — read when checking $var.field references in prompts to verify the attribute exists on the content type

---

Method summary (on request only)

After validation passes, do not volunteer a walkthrough of the method — no pipeline-flow diagram, no step-by-step breakdown of the pipes, no "here's what it does" narration. The platform renders the method visually, so a written summary here is redundant noise.

On a clean check, confirm in one line (e.g. "<name> is valid."). Never hide problems: still report every error, warning, and suggestion plainly — that is this skill's job.

Give the full summary (flow + steps) only when the user explicitly asks — "explain it", "summarize", "what does it do", "walk me through".