skill-improver
Pass
Audited by Gen Agent Trust Hub on Feb 17, 2026
Risk Level: SAFENO_CODE
Full Analysis
- Prompt Injection (SAFE): No evidence of instructions designed to bypass agent constraints or safety protocols. The skill's instructions are focused on its stated purpose of process reflection.
- Data Exposure & Exfiltration (SAFE): No access to sensitive files or unauthorized network communication was detected.
- Indirect Prompt Injection (LOW): The skill possesses an attack surface where malicious instructions could be embedded in the data being 'reflected' upon. (1) Ingestion point: User-provided session context and process descriptions. (2) Boundary markers: Present via 'Improvement Principles' requiring evidence and cost/benefit analysis. (3) Capability inventory: Potential modification of agent skills via the 'skill-creator' tool. (4) Sanitization: Relies on agent-level reasoning and manual user oversight. This creates a low-risk surface for potential behavioral manipulation.
- No Code (SAFE): The skill consists entirely of markdown documentation and does not include any executable scripts, binary files, or configuration files that trigger command execution.
Audit Metadata