maestro-roadmap

Pass

Audited by Gen Agent Trust Hub on Apr 17, 2026

Risk Level: SAFECOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
  • [COMMAND_EXECUTION]: The skill utilizes the Bash tool for managing session directories and logging discoveries. The discovery board protocol uses shell redirection to append data from sub-agents to a file, which poses a risk of command injection if the sub-agent output is manipulated.
  • [PROMPT_INJECTION]: The skill ingests untrusted data from arguments and external files to build instructions for its CSV-based agent orchestration.
  • Ingestion points: Data is pulled from command arguments, brainstorm session exports, and local file references.
  • Boundary markers: The skill lacks explicit protective markers or delimiters when interpolating external content into agent instructions.
  • Capability inventory: The skill and its sub-agents have access to powerful tools like Bash, Write, and Edit.
  • Sanitization: Path-level sanitization is present for folder names, but no content sanitization is performed on the instructions provided to the agents.
Audit Metadata
Risk Level
SAFE
Analyzed
Apr 17, 2026, 01:12 AM
Security Audit — agent-trust-hub — maestro-roadmap