skills/ghuron/skills/epson/Gen Agent Trust Hub

epson

Pass

Audited by Gen Agent Trust Hub on Oct 13, 2026

Risk Level: SAFECOMMAND_EXECUTIONDATA_EXFILTRATIONINDIRECT_PROMPT_INJECTION
Full Analysis
  • [COMMAND_EXECUTION]: Runs a Node.js script (epson.mjs) to query the printer's status using the command node epson.mjs ink.
  • [DATA_EXFILTRATION]: The script performs network requests to a user-defined printer host (EPSON_HOST) and explicitly disables SSL certificate validation.
  • Evidence: The script initializes an HTTPS agent with rejectUnauthorized: false in epson.mjs. This facilitates communication with IoT devices using self-signed certificates but removes standard protection against man-in-the-middle (MITM) attacks on the local network.
  • [INDIRECT_PROMPT_INJECTION]: The skill ingests and parses HTML from a network-connected printer, creating a surface for potential instruction injection via crafted printer responses.
  • Ingestion points: Fetches HTML status pages from the IP or hostname specified by the EPSON_HOST environment variable in epson.mjs.
  • Boundary markers: Absent; the parsed data is provided to the agent as a structured JSON array without explicit instructions to ignore embedded content.
  • Capability inventory: The script is limited to reading printer status via HTTP GET and outputting the resulting JSON to the standard output.
  • Sanitization: Employs regular expressions in the parseInkLevels function to isolate ink level and color data from the printer's HTML response, providing basic structural validation but no sanitization of the text values.
Audit Metadata
Risk Level
SAFE
Analyzed
Oct 13, 2026, 09:30 AM