supabase-audit-tables-read
Pass
Audited by Gen Agent Trust Hub on Sep 14, 2026
Risk Level: SAFECOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTIONEXTERNAL_DOWNLOADS
Full Analysis
- [COMMAND_EXECUTION]: The skill instructs the agent to execute shell commands using
curlto interact with the Supabase REST API. These commands are used to test data access and are restricted to standard GET/HEAD requests for auditing purposes. - [EXTERNAL_DOWNLOADS]: The skill interacts with external Supabase project URLs (e.g.,
https://[project].supabase.co/rest/v1/). This is a core part of its auditing functionality and uses the user-provided project configuration. - [INDIRECT_PROMPT_INJECTION]: The skill possesses a surface for indirect prompt injection as it ingests untrusted data from external database tables.
- Ingestion points: Data is retrieved from Supabase tables via
curlrequests (documented inSKILL.md). - Boundary markers: The instructions do not specify explicit delimiters for the incoming table data, though they provide guidance on redacting PII in output.
- Capability inventory: The skill can write results to local files (
.sb-pentest-context.json,.sb-pentest-audit.log) and execute network requests viacurl. - Sanitization: The skill includes instructions to redact PII and secrets before logging sample data to evidence files.
Audit Metadata