The Agent Skills Directory

Security Best Practices Incorporation: The skill proactively addresses security risks by explicitly instructing the agent to avoid dangerous Python functions such as eval, exec, and pickle when processing untrusted data (Category 10).
Secret Management Guidelines: Section 6 of the core checklist specifically mandates that secrets should never be hard-coded, which aligns with industry-standard security practices for credential management (Category 2).
Secure Data Loading Patterns: In the provided reference examples, the skill demonstrates the use of yaml.safe_load() rather than the unsafe yaml.load(), promoting secure deserialization patterns.
Robust Error Handling: The instructions emphasize raising specific exceptions and providing context rather than using silent failures or broad except: blocks, which improves the auditability and stability of the code.

python-foundational