founder-intelligence

Pass

Audited by Gen Agent Trust Hub on Nov 16, 2026

Risk Level: SAFECOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
  • [COMMAND_EXECUTION]: The skill instructs the agent to execute a local Node.js script (scripts/lens-router.mjs) using user-provided text as a command-line argument to perform deterministic routing of business decisions to specific analysis frameworks. While the script itself contains only safe regular expression logic, the execution of subprocesses with user input is a security-sensitive operation.
  • [INDIRECT_PROMPT_INJECTION]: The skill features a vulnerability surface where untrusted user input (the 'decision statement') is directly interpolated into a shell command template.
  • Ingestion points: The user's 'decision or obstacle' is captured as a required input in the SKILL.md workflow.
  • Boundary markers: The instructions wrap the input in double quotes ("decision statement") but fail to include explicit directives for the agent to sanitize or escape shell metacharacters (e.g., ;, &, |, `, $()).
  • Capability inventory: The skill has the capability to execute local scripts and read/write configuration files to the dedicated ${CLAUDE_PLUGIN_DATA} directory.
  • Sanitization: There is no evidence of input validation or sanitization within the SKILL.md instructions or the lens-router.mjs script to prevent command injection.
Audit Metadata
Risk Level
SAFE
Analyzed
Nov 16, 2026, 08:01 AM