vector-db-audit

Pass

Audited by Gen Agent Trust Hub on Jun 12, 2026

Risk Level: SAFECOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
  • [SAFE]: No obfuscation, unauthorized network access, or credential exposure was detected. The skill's operations are consistent with its stated purpose.
  • [COMMAND_EXECUTION]: The skill instructs the agent to execute a local Python script (scripts/audit_vector.py) via the Bash tool to perform the audit logic. This is a standard execution pattern for local project tools.
  • [PROMPT_INJECTION]: The skill involves a data ingestion surface where external content is processed. While this provides a surface for indirect prompt injection, it is necessary for the skill's primary function.
  • Ingestion points: The skill reads project documentation manifests and document chunks from ChromaDB collections.
  • Boundary markers: None explicitly defined in the instructions; processing is deferred to an external script.
  • Capability inventory: The skill is permitted to use Bash for execution and Write for generating CSV and report files.
  • Sanitization: No specific sanitization or validation of the manifest or database content is mentioned in the instruction file.
Audit Metadata
Risk Level
SAFE
Analyzed
Jun 12, 2026, 08:07 PM
Security Audit — agent-trust-hub — vector-db-audit