portravo-prepare-takeoff
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 plans in PDF format using the
attach_plan_pdftool. - Ingestion points: Untrusted data enters the agent's context via user-provided PDF files in
SKILL.md(instructions forattach_plan_pdf). - Capability inventory: The skill possesses the ability to create projects (
create_project), initiate tasks involving credit usage (run_takeoff), and retrieve detailed takeoff data (get_takeoff). - Boundary markers: The instructions lack explicit boundary markers or directions to ignore instructions potentially embedded within the PDF content.
- Sanitization: There is no mention of sanitization, filtering, or validation of content extracted from the PDF files prior to agent processing.
Audit Metadata