guv3-add-tool

Pass

Audited by Gen Agent Trust Hub on Mar 4, 2026

Risk Level: SAFEPROMPT_INJECTION
Full Analysis
  • [PROMPT_INJECTION]: The skill describes a template for creating tools that ingest untrusted data from an LLM context, which constitutes an indirect prompt injection attack surface.
  • Ingestion points: The your_tool_name function template in SKILL.md accepts parameters (e.g., param1, param2) that are populated by the LLM based on user input.
  • Boundary markers: The provided code examples do not include boundary markers or delimiters (such as XML tags or triple quotes) to isolate user-provided data from tool logic.
  • Capability inventory: The skill outlines significant capabilities including database modifications (update_one, insert_one in gu/db/connections/mongo) and communication actions (send_whatsapp_message, send_notification_to_owner in gu/core/notification_functions).
  • Sanitization: The templates demonstrate basic error handling and type checking via Pydantic schemas, but do not provide examples of content sanitization or instruction filtering for the ingested data.
Audit Metadata
Risk Level
SAFE
Analyzed
Mar 4, 2026, 04:12 AM