G-Pilot Fleet Operations

Pass

Audited by Gen Agent Trust Hub on Feb 21, 2026

Risk Level: SAFENO_CODEPROMPT_INJECTION
Full Analysis
  • [PROMPT_INJECTION] (LOW): Indirect Prompt Injection Surface. The protocol describes a 'Mission Lifecycle' that processes untrusted external data to drive agent actions.
  • Ingestion points: User requests received via the Dashboard are processed by an 'Architect Node' (SKILL.md).
  • Boundary markers: The documentation does not specify the use of delimiters or warnings to prevent the model from obeying instructions embedded within the user data.
  • Capability inventory: The 'Executor Node' has the capability to trigger Google Slides, Sheets, and Web Intel tools via a Google Auth bridge (SKILL.md).
  • Sanitization: There is no mention of input sanitization, schema validation, or escaping of the 'Mission SPEC' before tool execution.
  • [NO_CODE] (SAFE): No executable scripts, binaries, or configuration files were provided in the skill. The analysis is based solely on the logic described in the documentation.
Audit Metadata
Risk Level
SAFE
Analyzed
Feb 21, 2026, 01:36 AM