The Agent Skills Directory

[REMOTE_CODE_EXECUTION] (LOW): In async.md, a code example for loading models includes trust_remote_code=True. This parameter allows the execution of arbitrary code from the model's source repository, which is a significant risk if the repository is not strictly verified.
[REMOTE_CODE_EXECUTION] (LOW): In patterns.md, a code example demonstrates loading YAML configuration using yaml_instance.load(f). Recommending the load() method instead of safe_load() or a restricted loader is a known security anti-pattern that can enable arbitrary code execution through malicious YAML payloads.
[SAFE] (SAFE): The automated scanner alert regarding logger.info is a false positive; the scanner appears to have incorrectly identified a standard Python logging method call as a malicious URL.

python-skills