The Agent Skills Directory

[SAFE]: A comprehensive analysis of the skill's instructions and Python script revealed no security threats. The skill operates locally on data files and performs deterministic schema transformations.
[DATA_EXFILTRATION]: The skill does not utilize any networking libraries. All operations are confined to the local filesystem within the provided workspace path.
[COMMAND_EXECUTION]: File operations are limited to reading JSONL, YAML, and BibTeX files and writing reports or updating JSONL files in place. There is no evidence of subprocess spawning or arbitrary command execution.
[REMOTE_CODE_EXECUTION]: The script does not download external code or use dynamic execution functions like eval() or exec(). It relies on standard library JSON and YAML parsing.
[PROMPT_INJECTION]: No instructions were found that attempt to bypass safety guidelines or override agent behavior. The metadata enforces a 'NO PROSE' policy to ensure deterministic output.

schema-normalizer