gemini

Pass

Audited by Gen Agent Trust Hub on Feb 27, 2026

Risk Level: SAFEPROMPT_INJECTIONCOMMAND_EXECUTION
Full Analysis
  • [PROMPT_INJECTION]: The skill is susceptible to indirect prompt injection because it processes untrusted external data such as entire codebases and documentation sets.
  • Ingestion points: Data enters the context via the gemini command when analyzing directories or files specified in the prompt or via include-directories (found in SKILL.md).
  • Boundary markers: Absent. No delimiters or safety warnings are suggested for input data to prevent the LLM from following instructions embedded in the code.
  • Capability inventory: Uses gemini CLI with --approval-mode yolo which auto-approves all internal tools, which may include file system modification or further command execution (found in SKILL.md).
  • Sanitization: Absent. No sanitization or validation of processed codebases is performed before execution.
  • [COMMAND_EXECUTION]: The skill instructs the agent to execute multiple shell commands including gemini, ps, pkill, and timeout to perform analysis and manage process lifecycles.
Audit Metadata
Risk Level
SAFE
Analyzed
Feb 27, 2026, 02:37 PM