eda-research
Pass
Audited by Gen Agent Trust Hub on Sep 19, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONEXTERNAL_DOWNLOADS
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill is susceptible to indirect prompt injection because it processes data from external web searches and component datasheets to generate project documentation and BOM files.
- Ingestion points: The skill uses
WebSearchandWebFetchto gather information from external websites and component datasheets as defined in the 'Analyze Candidates' workflow inSKILL.md. - Boundary markers: There are no explicit instructions for the agent to treat external content as untrusted or to ignore instructions embedded within the processed datasheets.
- Capability inventory: The agent has the capability to write to local project files (
docs/component-selections.md,docs/bom-draft.json) and execute library modification commands viamcp__jlc__library_fix. - Sanitization: No sanitization or validation mechanisms are mentioned for the data extracted from external PDFs or web content before it is used to populate documentation.
- [EXTERNAL_DOWNLOADS]: The skill downloads component datasheets and reference materials from manufacturer websites and distributors using
WebFetch. This is a standard and necessary component of the skill's stated workflow for EDA research.
Audit Metadata