dbk-python-style
Pass
Audited by Gen Agent Trust Hub on Oct 13, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [SAFE]: The skill defines legitimate development workflows and coding standards. It encourages the use of established, reputable tools and libraries such as
uv,pathlib,rich, andlogurufor project management and logging. No evidence of obfuscation, unauthorized data access, or remote code execution was found. - [INDIRECT_PROMPT_INJECTION]: The skill involves processing and refactoring user-provided Python source code, which represents a potential surface for indirect prompt injection common to coding assistant tools.
- Ingestion points: User-provided Python files (.py) and natural language queries as described in the metadata and evals/trigger-queries.json.
- Boundary markers: The skill does not define specific boundary markers for untrusted code, but it does include a safeguard instruction requiring the agent to ask the user before performing large-scale refactoring.
- Capability inventory: The agent is instructed to modify local files, manage dependencies, and execute environment management tasks via the
uvtool (SKILL.md). - Sanitization: No explicit sanitization or filtering of input code content is specified within the skill instructions.
Audit Metadata