concept-classifier

Warn

Audited by Gen Agent Trust Hub on Feb 16, 2026

Risk Level: MEDIUMPROMPT_INJECTIONEXTERNAL_DOWNLOADS
Full Analysis
  • Indirect Prompt Injection (MEDIUM): The skill creates an attack surface where untrusted data from a 'data.json' file is incorporated into generated '.js' and '.html' files. Ingestion points: 'scenarios' and 'explanation' fields in data.json. Boundary markers: Absent; there are no delimiters or instructions to treat data as non-executable. Capability inventory: The skill generates and writes web-executable files to the filesystem. Sanitization: Absent; the skill does not include logic to sanitize input before file generation.
  • External Downloads (MEDIUM): The technical notes specify 'p5.js 1.11.10'. As this version is not a standard release (1.11.0 is the current latest), it represents an unverifiable dependency that could lead to loading compromised code if the agent or environment attempts to resolve it.
Audit Metadata
Risk Level
MEDIUM
Analyzed
Feb 16, 2026, 01:36 PM