The Agent Skills Directory

[SAFE]: The core philosophy establishes a 'Security Priority' requiring the use of read_file to confirm context before performing any write_file operation, which prevents blind or unsafe code edits.
[SAFE]: The skill includes explicit security quality gates that mandate input validation, proper exception handling, and the prevention of sensitive information disclosure in generated code.
[SAFE]: A robust 'Conflict Protocol' and 'Blocking Report' system ensure the agent stops and seeks human intervention when instructions are unclear or conflict with existing code, preventing unintended or risky changes.

implementation-engineer