wordpress-plugin-fundamentals
Pass
Audited by Gen Agent Trust Hub on Feb 15, 2026
Risk Level: LOWEXTERNAL_DOWNLOADSPROMPT_INJECTIONDATA_EXFILTRATION
Full Analysis
- EXTERNAL_DOWNLOADS (LOW): The installation instructions in README.md and metadata.json reference
wp-coding-standards/wpcsandphpunit/phpunitvia Composer. - Evidence:
composer require --dev wp-coding-standards/wpcs:"^3.0"andphpunit/phpunit:"^9.6". - Context: These are standard, reputable development tools. Per [TRUST-SCOPE-RULE], these are downgraded to LOW/INFO as they target trusted registries.
- PROMPT_INJECTION (LOW): The skill provides templates for code generation, which inherently creates an indirect prompt injection surface when used by an agent to process untrusted user requests.
- Ingestion points: User-provided plugin requirements and feature descriptions.
- Boundary markers: Absent in the raw code templates, though the skill provides narrative context.
- Capability inventory: The agent generates PHP code and modifies files (HIGH capability tier).
- Sanitization: The skill provides exceptional mitigation by including mandatory security sections (sanitization, nonces, prepared statements) and requiring WPCS compliance. The severity is LOW because the skill's primary purpose is to enforce security standards.
- DATA_EXFILTRATION (LOW): The README.md contains a hardcoded local file path from the author's environment.
- Evidence:
/Users/masa/Projects/claude-mpm-skills/docs/research/wordpress-development-ecosystem-2025-01-30.md. - Context: This is a negligible metadata leak regarding the author's project structure and does not facilitate data exfiltration from the user's system.
Audit Metadata