harvest-structured

Pass

Audited by Gen Agent Trust Hub on Jun 22, 2026

Risk Level: SAFEPROMPT_INJECTIONEXTERNAL_DOWNLOADS
Full Analysis
  • [PROMPT_INJECTION]: The skill is susceptible to indirect prompt injection attacks.
  • Ingestion points: The skill uses WebFetch to retrieve content from arbitrary URLs provided by the user at runtime (SKILL.md).
  • Boundary markers: The instructions do not specify any boundary markers (such as XML tags or unique delimiters) or provide explicit directions for the agent to ignore instructions embedded within the fetched HTML content.
  • Capability inventory: The skill's execution environment allows for the use of Bash (shell command execution), Write (file system modification), and WebFetch (network requests).
  • Sanitization: There are no provisions for sanitizing, filtering, or validating the external content before it is processed by the agent's logic.
  • [EXTERNAL_DOWNLOADS]: The skill is designed to download content from external, untrusted sources via the WebFetch tool to perform its primary function of data extraction.
Audit Metadata
Risk Level
SAFE
Analyzed
Jun 22, 2026, 07:46 AM
Security Audit — agent-trust-hub — harvest-structured