supabase
Pass
Audited by Gen Agent Trust Hub on Feb 17, 2026
Risk Level: SAFE
Full Analysis
- Prompt Injection (SAFE): No instructions designed to override AI safety or bypass instructions were found. The skill contains a safety warning in the documentation body to prevent accidental command execution.
- Data Exposure & Exfiltration (SAFE): No hardcoded credentials, API keys, or access to sensitive local file paths (like .ssh or .aws) were identified. There is no code capable of making network requests.
- Obfuscation (SAFE): The documentation and metadata are written in clear, human-readable text. No Base64, zero-width characters, or homoglyph-based obfuscation techniques are present.
- Unverifiable Dependencies & Remote Code Execution (SAFE): The skill does not include any dependency installation commands (npm/pip) or patterns involving the download and execution of remote scripts.
- Indirect Prompt Injection (SAFE): While the skill involves the agent processing external data, the source (Supabase official documentation) is considered a trusted source, and the content is static.
- Ingestion points: Files in the 'references/' directory.
- Boundary markers: The 'Security Notice' in SKILL.md serves as a delimiter-like instruction for the agent.
- Capability inventory: No execution capabilities (subprocess, eval, or network) exist in the skill's file set.
- Sanitization: Not applicable as the content is static documentation.
- Dynamic Execution (SAFE): No runtime code generation, compilation, or unsafe deserialization (e.g., pickle/yaml) was found.
Audit Metadata