jevgraph
Pass
Audited by Gen Agent Trust Hub on Sep 21, 2026
Risk Level: SAFEEXTERNAL_DOWNLOADSCOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [EXTERNAL_DOWNLOADS]: The skill instructs the agent to download and execute code from the author's GitHub repository (github:Emlembow/jevgraph) using npx or npm install --global.
- [COMMAND_EXECUTION]: The skill relies on executing shell commands (jevgraph, node bin/jevgraph.js) to perform its primary functions.
- [INDIRECT_PROMPT_INJECTION]: The skill processes untrusted external data from local snapshots and Notion pages. 1. Ingestion points: Reads data from JSON/Markdown snapshots and Notion API content (SKILL.md, references/schema.md). 2. Boundary markers: The skill emphasizes evidence-preserving audits and reporting evidence from page content, which implies a degree of data integrity checking, though explicit delimiters for prompt protection are not defined. 3. Capability inventory: The agent executes shell commands (jevgraph) and performs semantic searches which could be influenced by malicious content in the processed snapshots. 4. Sanitization: The skill mentions redacted diagnostics for configuration checks but does not detail specific sanitization for processed graph data.
Audit Metadata