humanizer-he

Pass

Audited by Gen Agent Trust Hub on Oct 5, 2026

Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
  • [INDIRECT_PROMPT_INJECTION]: The skill is designed to ingest and process Hebrew text from external sources, which creates a potential attack surface for indirect prompt injection. A malicious actor could provide text containing embedded instructions intended to override the agent's primary task or manipulate its tools. Ingestion points: According to the 'Usage Notes', the skill accepts input through pasted text or via file paths. Boundary markers: The instructions do not define specific delimiters or instructions to treat input text as data only, lacking explicit boundaries to separate user input from system-level instructions. Capability inventory: The agent is authorized to use 'Read', 'Write', 'Edit', and 'AskUserQuestion' tools, providing a surface for filesystem modification if an injection is successful. Sanitization: There are no procedural steps or instructions provided to sanitize, filter, or validate the input Hebrew text for embedded commands prior to processing.
Audit Metadata
Risk Level
SAFE
Analyzed
Oct 5, 2026, 08:40 AM