training-path

Pass

Audited by Gen Agent Trust Hub on Mar 23, 2026

Risk Level: SAFEPROMPT_INJECTIONCOMMAND_EXECUTION
Full Analysis
  • [PROMPT_INJECTION]: The skill creates an indirect prompt injection surface by processing untrusted external data. Ingestion points: The agent reads codebase manifests, data models, and learner profiles from the environment (SKILL.md). Boundary markers: No explicit delimiters or boundary markers are defined to isolate processed content from system instructions. Capability inventory: The skill has permission to read the filesystem and write analysis reports to the local directory (SKILL.md). Sanitization: The instructions include a manual safety check requiring the agent to exclude personally identifiable information (PII) from its output.
  • [COMMAND_EXECUTION]: The skill performs local file system operations to read the project codebase for its analysis phases. It writes findings to the docs/ folder and appends execution metadata to the platform-specific project directory in ~/.claude/projects/.
Audit Metadata
Risk Level
SAFE
Analyzed
Mar 23, 2026, 10:58 AM
Security Audit — agent-trust-hub — training-path