ISA

Pass

Audited by Gen Agent Trust Hub on May 2, 2026

Risk Level: SAFECOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
  • [COMMAND_EXECUTION]: The skill instructs the agent to execute curl commands to send JSON-formatted status notifications to a local server (localhost:31337) whenever a workflow is initiated. This is a local-only telemetry hook for the author's environment and does not involve external network calls.
  • [PROMPT_INJECTION]: The Seed workflow ingests data from local repository files such as README.md, package.json, and git logs to generate initial ISA documentation. This introduces a surface for indirect prompt injection, as malicious instructions within those untrusted files could influence the generated draft. This risk is localized to the resulting document.
  • [SAFE]: The skill operates within expected directories and uses standard agent capabilities to manage files. All identified patterns are operational features of the Personal AI (PAI) framework for which the skill was designed.
Audit Metadata
Risk Level
SAFE
Analyzed
May 2, 2026, 01:03 AM