plan-marshall-plugin

Pass

Audited by Gen Agent Trust Hub on Jul 11, 2026

Risk Level: SAFE
Full Analysis
  • [SAFE]: The skill acts as a central discovery hub for the Plan Marshall platform, identifying build system markers and delegating discovery to specialized internal scripts. All operations are confined to build file identification and metadata extraction.
  • [PROMPT_INJECTION]: The skill processes untrusted project files such as pom.xml, package.json, and pyproject.toml during the module discovery phase. This is an inherent property of build discovery tools.
  • Ingestion points: extension.py (discover_modules) and _discover_nested_descriptors methods.
  • Boundary markers: Handled by internal platform delegation to sibling discovery scripts.
  • Capability inventory: File system read operations (marker detection) and orchestration of module discovery.
  • Sanitization: Validation is performed by the delegated discovery scripts for each build system.
  • [REMOTE_CODE_EXECUTION]: The Python extension uses dynamic imports to load discovery logic from sibling skill directories. This represents a standard architectural pattern for the platform's plugin system and involves internal vendor scripts rather than untrusted remote code.
Audit Metadata
Risk Level
SAFE
Analyzed
Jul 11, 2026, 07:13 AM
Security Audit — agent-trust-hub — plan-marshall-plugin