sn-research-report
Pass
Audited by Gen Agent Trust Hub on Apr 29, 2026
Risk Level: SAFECOMMAND_EXECUTIONEXTERNAL_DOWNLOADSPROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill instructs the agent to execute a shell command
python3 <sn-image-base>/scripts/sn_agent_runner.py sn-image-generateto generate images. This is a functional requirement for its 'AI image' feature and targets a vendor-prefixed dependency (sn-image-base). - [EXTERNAL_DOWNLOADS]: The skill defines functional dependencies on other skills (
sn-deep-researchandsn-image-base). It expects these skills to be present in the environment to function correctly. - [PROMPT_INJECTION]: The skill exhibits an indirect prompt injection surface. It ingests untrusted data from multiple sources (research notes, sub-reports, and user-provided drafts) and uses this content to formulate reports and image generation prompts.
- Ingestion points: Reads files such as
synthesis.md,sub_reports/*.md, and user-supplied text snippets. - Boundary markers: The instructions do not define explicit delimiters or 'ignore' instructions for the content being processed.
- Capability inventory: The skill has the capability to write files (
report.md,visual_plan.md) and execute shell commands (python3viasn-image-base). - Sanitization: There is no mention of sanitizing or escaping the content retrieved from research materials before it is interpolated into subsequent prompts or command arguments.
Audit Metadata