NYC

bv

Pass

Audited by Gen Agent Trust Hub on Feb 17, 2026

Risk Level: SAFE
Full Analysis
  • Indirect Prompt Injection (LOW): This skill parses '.beads/beads.jsonl' to provide task recommendations. Malicious instructions could be placed in bead titles or descriptions by an attacker to manipulate agent logic. \n
  • Ingestion points: '.beads/beads.jsonl' processed by the 'bv' command. \n
  • Boundary markers: Structured JSON output is provided via '--robot-*' flags, though field content remains untrusted. \n
  • Capability inventory: Local execution of 'bv', 'bd', and 'jq' CLI tools. \n
  • Sanitization: No sanitization of project file content is described in the documentation. \n- Command Execution (SAFE): The documented commands are restricted to local analysis of the project dependency graph using the 'bv' utility and do not show signs of malicious command injection.
Audit Metadata
Risk Level
SAFE
Analyzed
Feb 17, 2026, 05:33 PM