sql-review
Pass
Audited by Gen Agent Trust Hub on Sep 20, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONCOMMAND_EXECUTIONDYNAMIC_EXECUTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill processes untrusted SQL code from the repository or user input to perform its analysis. * Ingestion points: SQL content read from file paths, git diff output, and direct conversation input. * Capability inventory: The skill has access to shell execution (bash), file reading (read), and filesystem globbing (glob). * Boundary markers: No explicit instructions are provided to the agent to treat the SQL content as untrusted data or to ignore instructions embedded within the SQL. * Sanitization: While the skill checks for SQL injection, it does not explicitly address prompt injection risks that could affect the AI agent's behavior.
- [COMMAND_EXECUTION]: The skill uses the bash tool to interact with git and the altimate-dbt CLI.
- [DYNAMIC_EXECUTION]: The workflow involves interpolating a model name into a shell command:
altimate-dbt compile --model <name>. If the model name (derived from file names or user input) contains shell control characters, it could lead to arbitrary command execution.
Audit Metadata