pm
Pass
Audited by Gen Agent Trust Hub on Oct 17, 2026
Risk Level: SAFEEXTERNAL_DOWNLOADSCOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [EXTERNAL_DOWNLOADS]: Performs a version check against the author's GitHub repository to detect updates.
- [COMMAND_EXECUTION]: Executes shell commands to update its own software components and uses
pbcopyto move generated content to the system clipboard. All update actions require user consent. - [INDIRECT_PROMPT_INJECTION]: Ingests potentially untrusted data from external code repositories.
- Ingestion points: External repository data accessed via fan-out subagents and initial user input.
- Boundary markers: Uses specific formatting and labeling to distinguish user-provided content.
- Capability inventory: Has access to shell commands and clipboard utilities.
- Sanitization: Employs a 'Refine gate' that forces human review of all structured outputs before they are processed further.
Audit Metadata