flux-query
Pass
Audited by Gen Agent Trust Hub on Jul 1, 2026
Risk Level: SAFECREDENTIALS_UNSAFEPROMPT_INJECTION
Full Analysis
- [CREDENTIALS_UNSAFE]: The instructions in
SKILL.md(Step 0) direct the agent to "Check for connection strings to identify the engine." Connection strings frequently contain sensitive plaintext credentials such as database passwords and authentication tokens. Accessing these values brings them into the agent's context, where they could be inadvertently exposed. - [PROMPT_INJECTION]: The skill is susceptible to indirect prompt injection (Category 8) because it processes untrusted data from the codebase and uses it to drive agent actions.
- Ingestion points: The agent reads SQL queries, ORM code, and execution plan outputs from the local environment (
SKILL.md, Steps 0 and 1). - Boundary markers: The skill does not provide instructions to wrap ingested data in delimiters or to ignore potential instructions embedded within SQL comments.
- Capability inventory: The skill is authorized to use high-privilege tools including
Bash,Write, andEdit(SKILL.md, allowed-tools). - Sanitization: No validation or sanitization is performed on the ingested content before the agent analyzes or acts upon it.
Audit Metadata