cut-ready
Pass
Audited by Gen Agent Trust Hub on Aug 22, 2026
Risk Level: SAFE
Full Analysis
- [EXTERNAL_DOWNLOADS]: The skill references documentation, design guidelines, and manufacturer safety warnings from well-known fabrication services and standards organizations, including SendCutSend, Roland DG, StickerApp, Sticker Mule, ISO, ASME, W3C, Adobe, and Epilog Laser. These references are used for technical guidance and safety information.
- [INDIRECT_PROMPT_INJECTION]: The skill defines workflows for ingesting and processing untrusted external files such as CAD geometry, vector artwork, and photographs.
- Ingestion points: The skill ingests physical job classifications and dimensions in
SKILL.md(Workflow Step 1) and the "Minimum manufacturing packet" inreferences/intake-and-revision.md. - Boundary markers: The skill lacks explicit boundary markers or delimiters to separate untrusted user data from agent instructions during prompt interpolation.
- Capability inventory: The skill is capable of generating and exporting various manufacturing file formats (SVG, PDF, DXF, STEP, AI, EPS, DWG) and requires capabilities for file system writes.
- Sanitization: The skill mandates extensive validation gates to mitigate risks from malformed or malicious data, including technical preflight checks in
references/2d-preflight.mdandreferences/3d-preflight.md, visual/semantic review inSKILL.md, and vendor-specific preview verification inreferences/vendor-handoff.md.
Audit Metadata