capital-allocation

Pass

Audited by Gen Agent Trust Hub on Jun 25, 2026

Risk Level: SAFEPROMPT_INJECTIONCOMMAND_EXECUTION
Full Analysis
  • [PROMPT_INJECTION]: The skill exhibits an indirect prompt injection surface by processing untrusted data from a JSON source without sufficient sanitization or boundary markers.
  • Ingestion points: The scripts/capital_allocate.py script ingests data via the load function from a user-specified file path or stdin.
  • Boundary markers: There are no explicit delimiters or instructions provided to the agent to treat the processed data as untrusted or to ignore embedded instructions.
  • Capability inventory: The skill performs local file reading and executes a Python script for deterministic sorting and calculation.
  • Sanitization: While the script uses standard JSON parsing (json.loads), it does not sanitize or escape the content of initiative names or other fields before they are output to the agent.
  • [COMMAND_EXECUTION]: The skill relies on the execution of a local Python script (scripts/capital_allocate.py) to perform its primary function. This is documented and uses only standard library modules, representing intended functionality rather than a malicious vector.
Audit Metadata
Risk Level
SAFE
Analyzed
Jun 25, 2026, 09:20 AM
Security Audit — agent-trust-hub — capital-allocation