supabase
Pass
Audited by Gen Agent Trust Hub on Oct 13, 2026
Risk Level: SAFECOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTIONDYNAMIC_EXECUTION
Full Analysis
- [COMMAND_EXECUTION]: The skill uses shell commands like
curl,grep,cut, andsedto interact with the Supabase Management and PostgREST APIs. These commands are used to manage environment variables and perform database operations on official Supabase domains (supabase.coandsupabase.com). - [INDIRECT_PROMPT_INJECTION]: The skill ingests data from external database sources through API responses and processes it using Python. While this constitutes an attack surface if the database content is maliciously crafted by an external party, the risk is mitigated by the skill's focus on structured data analysis (e.g., counting records, calculating revenue) rather than interpreting the data as instructions.
- Ingestion points: API responses from
curlcommands inSKILL.md. - Boundary markers: None explicitly defined for API data.
- Capability inventory: Subprocess execution via
Bashtool, file system access viaRead/Writetools. - Sanitization: Standard JSON parsing using the
jsonPython module. - [DYNAMIC_EXECUTION]: Uses inline Python scripts (
python3 -c) to parse JSON output from API calls. These scripts are generated from static templates to perform specific calculations like MRR (Monthly Recurring Revenue) and conversion funnels, which is a standard method for data processing in this context.
Audit Metadata