python-best-practices-type-safety
Pass
Audited by Gen Agent Trust Hub on Feb 24, 2026
Risk Level: SAFE
Full Analysis
- [PROMPT_INJECTION]: Analysis of the skill instructions and metadata revealed no attempts to bypass safety filters, override system instructions, or extract system prompts. The instructional language is strictly task-oriented.
- [DATA_EXFILTRATION]: No sensitive file paths, hardcoded credentials, or unauthorized network operations were detected. Data handling is restricted to reading and editing local Python source files and type-checker logs.
- [REMOTE_CODE_EXECUTION]: The skill uses standard, well-known development tools (uv, pyright). No patterns of downloading and executing untrusted remote scripts (e.g., curl|bash) were found.
- [COMMAND_EXECUTION]: Commands are restricted to running the pyright type checker and standard Python script execution. There are no signs of privilege escalation or unauthorized system modifications.
- [INDIRECT_PROMPT_INJECTION]: While the skill processes external Python source code (an ingestion surface), it employs Abstract Syntax Tree (AST) parsing for analysis rather than string evaluation, reducing the risk of code injection. The capability to edit files is limited to the user's specific project files as intended by the primary function of the skill.
- [DYNAMIC_EXECUTION]: No usage of unsafe deserialization (pickle) or dynamic code execution (eval/exec) on untrusted data was found. File modifications are performed using standard line-based editing.
Audit Metadata