blender-node-graph
Pass
Audited by Gen Agent Trust Hub on Oct 8, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill exposes an attack surface for indirect prompt injection by ingesting untrusted data from the Blender scene environment.
- Ingestion points: The scripts
scripts/list_all_node_graphs.py,scripts/get_compositor_node_tree.py, andscripts/list_compositor_nodes.pyretrieve node names, labels, and tree structures from the active Blender scene. - Boundary markers: None are implemented in the scripts to delimit Blender scene data from instructions.
- Capability inventory: The skill allows access to
BashandReadtools as defined in the frontmatter configuration. - Sanitization: There is no evidence of sanitization or filtering of the strings retrieved from the Blender scene state.
Audit Metadata