robot-expert

Warn

Audited by Gen Agent Trust Hub on Feb 16, 2026

Risk Level: MEDIUMPROMPT_INJECTION
Full Analysis
  • PROMPT_INJECTION (MEDIUM): The skill directs the agent to load and follow instructions from an external path: {{cookiecutter.primary_pipeline_name}}/.claude/skills/robot-expert/SKILL.md. 1. Ingestion points: External data enters the context via the Read tool from a file path containing a template variable. 2. Boundary markers: Absent. There are no delimiters or instructions to ignore embedded commands within the loaded file. 3. Capability inventory: The agent is primarily used for guidance and referencing patterns; however, the 'Follow the Guide' instruction (Step 3) effectively delegates control of the agent's logic to the external file. 4. Sanitization: None. The agent does not validate or filter the content of the loaded file before adopting it as the primary instruction set.
Audit Metadata
Risk Level
MEDIUM
Analyzed
Feb 16, 2026, 08:05 AM