ln-57-persistence-auditor
Pass
Audited by Gen Agent Trust Hub on Oct 5, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [SAFE]: The skill is a read-only audit tool for database persistence and runtime resource management. It includes explicit prohibitions against connecting to production environments or performing mutating operations, such as creating indexes, migrating schemas, or rewriting data.
- [INDIRECT_PROMPT_INJECTION]: The skill has a potential attack surface for indirect prompt injection based on the following evaluation: 1. Ingestion points: Reads repository manifests, models, migrations, query logs, metrics, traces, and profiles (SKILL.md). 2. Boundary markers: None explicitly defined for prompt isolation. 3. Capability inventory: Uses native file search, language server tools, and database diagnostic tools (e.g., EXPLAIN) on non-production targets. 4. Sanitization: Mandates redaction of sensitive parameters, credentials, and row data before recording results.
Audit Metadata