snyk-security-scan
Pass
Audited by Gen Agent Trust Hub on Sep 21, 2026
Risk Level: SAFEEXTERNAL_DOWNLOADSDATA_EXFILTRATIONINDIRECT_PROMPT_INJECTIONDYNAMIC_EXECUTIONCOMMAND_EXECUTIONREMOTE_CODE_EXECUTION
Full Analysis
- [EXTERNAL_DOWNLOADS]: The skill downloads the pinned Snyk CLI (version 1.1306.0) from the official npm registry using the npx utility.
- [DATA_EXFILTRATION]: The skill transmits project metadata and source code to Snyk's cloud infrastructure for analysis. This is a documented and core feature of the Snyk security platform.
- [DYNAMIC_EXECUTION]: The skill utilizes npx to dynamically load and execute the Snyk CLI to facilitate scanning and authentication tasks.
- [COMMAND_EXECUTION]: The skill executes shell commands via the Snyk CLI to interact with the file system, manage project trust, and perform security audits.
- [REMOTE_CODE_EXECUTION]: The use of npx facilitates the downloading and execution of the Snyk CLI from a remote repository.
- [INDIRECT_PROMPT_INJECTION]: The skill ingests and processes untrusted data from the local file system (source code, manifests, SBOMs) that could contain malicious instructions designed to influence the agent's interpretation of scanner results. 1. Ingestion points: Source code, package manifests, and SBOM files read during scans. 2. Boundary markers: Not explicitly defined in the prompt instructions for tool output parsing. 3. Capability inventory: File system access, network communication with Snyk APIs, and command execution. 4. Sanitization: No specific filtering of tool output is performed before the agent processes the findings.
Audit Metadata