assumption-mapping

Pass

Audited by Gen Agent Trust Hub on Feb 17, 2026

Risk Level: SAFEPROMPT_INJECTIONNO_CODE
Full Analysis
  • [PROMPT_INJECTION] (LOW): Indirect prompt injection surface detected in the assumption extraction and file-saving workflow. Evidence: 1. Ingestion points: User-provided phrases and conversation context during assumption extraction. 2. Boundary markers: Absent; the agent is not instructed to use delimiters or ignore instructions embedded in user input. 3. Capability inventory: File-write capability to 'design/03-assumption-mapping.md'. 4. Sanitization: Absent; user strings are interpolated directly into the output template.
  • [NO_CODE] (SAFE): The skill consists entirely of markdown documentation and logic; no Python, Node.js, or shell scripts are included or executed.
  • [SAFE] (SAFE): No patterns of data exfiltration, hardcoded credentials, persistence mechanisms, or unauthorized privilege escalation were found.
Audit Metadata
Risk Level
SAFE
Analyzed
Feb 17, 2026, 06:09 PM