The Agent Skills Directory

[COMMAND_EXECUTION]: The skill workflow involves executing local environment commands such as linting, type checking, and running unit tests to verify the behavior of refactored code. It specifically mentions running a local validation script named validate-yaml.py to check the integrity of configuration files.
[DATA_EXFILTRATION]: Method M-L1-04 (Characterization Test) suggests that the agent sample "production logs" to generate test inputs and outputs. While this is a standard engineering practice for stabilizing legacy code, users should be aware that ingesting raw production logs into the AI context may expose sensitive information (PII or credentials) contained within those logs.
[SAFE]: The skill implements significant security best practices, including explicit refusal instructions for out-of-scope tasks and mandatory human review before any code modifications are applied.

cs-refactor