archiving-databases

Pass

Audited by Gen Agent Trust Hub on Jun 16, 2026

Risk Level: SAFENO_CODE
Full Analysis
  • [SAFE]: No malicious patterns, such as prompt injection, data exfiltration, or obfuscation, were detected in the skill's instructions or metadata.
  • [NO_CODE]: The skill does not include any scripts, executables, or code files. It defines the archival process through high-level instructions and examples for the agent to follow using its existing toolset.
  • [PROMPT_INJECTION]: The skill manages data from external databases (PostgreSQL/MySQL), which represents a potential surface for indirect prompt injection if processed records contain adversarial instructions. However, the skill does not exhibit any patterns of bypassing safety protocols. 1. Ingestion points: External database records identified for archival. 2. Boundary markers: None explicitly specified. 3. Capability inventory: The skill is configured to use tools including Bash, Read, and Write. 4. Sanitization: No specific sanitization or validation logic is provided for the data being archived.
Audit Metadata
Risk Level
SAFE
Analyzed
Jun 16, 2026, 01:01 AM
Security Audit — agent-trust-hub — archiving-databases