validate-plan-file

Pass

Audited by Gen Agent Trust Hub on Jul 18, 2026

Risk Level: SAFEPROMPT_INJECTION
Full Analysis
  • [PROMPT_INJECTION]: The skill presents a surface for indirect prompt injection by reading and processing external JSON plan files. If these files are retrieved from untrusted sources, they could contain embedded instructions intended to manipulate the agent's logic in subsequent tasks.
  • Ingestion points: The skill reads files from the local filesystem using the Read tool (e.g., .bug-detection-plan.json, .security-scan-plan.json).
  • Boundary markers: There are no explicit delimiters or instructions provided to the agent to disregard natural language instructions that might be embedded within the JSON values.
  • Capability inventory: The skill uses the Read tool to access plan files and schema files, and performs JSON schema validation.
  • Sanitization: The skill does not implement sanitization or filtering of the JSON content beyond structural schema validation.
Audit Metadata
Risk Level
SAFE
Analyzed
Jul 18, 2026, 01:47 PM
Security Audit — agent-trust-hub — validate-plan-file