data-warehousing
Pass
Audited by Gen Agent Trust Hub on Apr 7, 2026
Risk Level: SAFECOMMAND_EXECUTIONEXTERNAL_DOWNLOADSPROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill contains instructions for the agent to execute a shell command (
ls) targeting several local directories (such as ~/.claude/skills/) to detect the presence of other installed skills. This behavior constitutes local environment reconnaissance. - [EXTERNAL_DOWNLOADS]: The documentation and agent instructions promote the installation of companion skills from an external repository (AbsolutelySkilled/AbsolutelySkilled) using
npx, which involves fetching and executing remote code. - [PROMPT_INJECTION]: The skill is vulnerable to indirect prompt injection because it ingests untrusted user requirements to generate SQL code without explicit safety boundaries or sanitization.
- Ingestion points: User-provided schema definitions and business logic descriptions enter the agent context.
- Boundary markers: None are defined in the skill instructions to separate user input from the generated SQL logic.
- Capability inventory: The agent is authorized to perform environment reconnaissance via shell commands.
- Sanitization: No validation, escaping, or filtering of user-supplied content is specified before interpolation into SQL templates.
Audit Metadata