skills/socketdev/sauce/socket-scan/Gen Agent Trust Hub

socket-scan

Pass

Audited by Gen Agent Trust Hub on Nov 14, 2026

Risk Level: SAFEINDIRECT_PROMPT_INJECTIONCOMMAND_EXECUTIONEXTERNAL_DOWNLOADS
Full Analysis
  • [INDIRECT_PROMPT_INJECTION]: The skill processes untrusted data from project manifest files and scan result JSON, creating a surface for indirect prompt injection attacks.
  • Ingestion points: The skill reads project configuration files (e.g., package.json, requirements.txt) and parses JSON output from socket scan and cdxgen (e.g., bom.json).
  • Boundary markers: Absent. There are no instructions to wrap untrusted data in specific delimiters or use strict system prompts to ignore embedded commands.
  • Capability inventory: The agent has the ability to execute shell commands via pnpm dlx, read local files, and communicate with the Socket API.
  • Sanitization: Absent. The skill does not define methods for escaping or validating fields like dependency names, version strings, or issue descriptions before processing them.
  • [COMMAND_EXECUTION]: The skill relies extensively on shell commands to check authentication, perform scans, and audit licenses.
  • Runs pnpm dlx socket for organization listing, scanning, and reachability analysis.
  • Executes pnpm dlx @cyclonedx/cdxgen as a fallback for SBOM generation.
  • Invokes various package manager audit tools such as npm audit, pnpm audit, and yarn audit.
  • [EXTERNAL_DOWNLOADS]: The skill downloads and executes CLI tools at runtime.
  • Fetches the socket tool, which is a vendor-owned resource for the author 'socketdev'.
  • Fetches @cyclonedx/cdxgen, which is a well-known industry standard tool for CycloneDX SBOM generation.
  • These operations target well-known public registries (npm) and trusted/well-known organizations.
Audit Metadata
Risk Level
SAFE
Analyzed
Nov 14, 2026, 07:55 PM