supabase-audit-buckets-public
Pass
Audited by Gen Agent Trust Hub on Sep 14, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill ingests metadata from an external environment (Supabase storage) which acts as an untrusted input surface.
- Ingestion points: Bucket names and object filenames fetched via Supabase API calls during the audit process (SKILL.md).
- Boundary markers: Absent. The instructions do not define delimiters or "ignore instructions" markers to isolate data retrieved from the storage environment.
- Capability inventory: The skill performs network operations to the Supabase API and possesses the capability to write findings to multiple local files, including .sb-pentest-context.json, .sb-pentest-audit.log, and the .sb-pentest-evidence/ directory (SKILL.md).
- Sanitization: Absent. The skill does not provide logic for validating or escaping strings retrieved from external storage before processing or logging them.
Audit Metadata