portravo-scope-questions
Pass
Audited by Gen Agent Trust Hub on Nov 8, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill processes external construction takeoff notes and project data which could potentially contain malicious instructions or unexpected content designed to influence the agent. The risk is assessed as low because the skill is restricted to generating text for human review and explicitly lacks the capability to send outgoing communications or modify data.
- Ingestion points: Raw takeoff notes and project metadata are ingested via the
get_takeofftool in SKILL.md. - Boundary markers: The instructions do not define explicit delimiters or headers to distinguish between system instructions and untrusted project data.
- Capability inventory: The skill uses tools to read project state (
list_projects,get_takeoff_status,get_takeoff) but does not possess tool access for file system writes, shell execution, or network exfiltration. - Sanitization: There is no mention of sanitizing or escaping the data retrieved from the Portravo API before it is processed.
Audit Metadata