cold-start-interview

Pass

Audited by Gen Agent Trust Hub on Jul 9, 2026

Risk Level: SAFEPROMPT_INJECTIONCOMMAND_EXECUTION
Full Analysis
  • [PROMPT_INJECTION]: The skill exhibits an indirect prompt injection surface (Category 8) by ingesting user responses and writing them to configuration files used by other accounting modules.\n
  • Ingestion points: User responses provided during the multi-part interview in SKILL.md.\n
  • Boundary markers: No specific delimiters or safety instructions are used when interpolating user strings into the firm-profile.md template.\n
  • Capability inventory: The skill possesses file-write capabilities (~/.claude/plugins/config/) and executes tool calls for integration verification.\n
  • Sanitization: The skill instructions do not specify any validation or sanitization for the ingested user content.\n- [COMMAND_EXECUTION]: The skill uses the --check-integrations flag to execute live tool calls to verify external connectors, which is consistent with its configuration purpose.
Audit Metadata
Risk Level
SAFE
Analyzed
Jul 9, 2026, 02:12 PM
Security Audit — agent-trust-hub — cold-start-interview