pencil-mcp-batch-get
Pass
Audited by Gen Agent Trust Hub on Sep 4, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill processes external data via the
filePathparameter, which creates a potential surface for indirect prompt injection if a design file contains malicious instructions. However, the skill provides robust mitigation instructions. - Ingestion points: The
batch_gettool accepts afilePathto search and read node information. - Boundary markers: The skill includes a 'CRITICAL' intent recognition section requiring the agent to verify that the user explicitly mentions 'Pencil' and that the context matches design tasks before proceeding.
- Capability inventory: The skill is limited to searching and reading node properties (
batch_get) within the design context. - Sanitization: The skill encourages 'Smart Traversal' and low search/read depths (<3), which limits the volume of data processed in a single context window.
Audit Metadata