The Agent Skills Directory

[COMMAND_EXECUTION]: The skill utilizes standard development tools and version control commands, including git, gh (GitHub CLI), and common test runners like npm, cargo, pytest, and go. These commands are used for their intended purposes within a development environment.
[SAFE]: The skill includes robust safety practices, such as requiring automated tests to pass before suggesting integration options and mandating explicit user confirmation before performing destructive actions like discarding a branch.
[PROMPT_INJECTION]: No patterns of instruction override, safety filter bypass, or role-play injections were detected in the instructions.
[DATA_EXFILTRATION]: No unauthorized data access or external transmission patterns were identified. Network activity is limited to standard Git operations (pull/push) and PR creation.
[OBFUSCATION]: The content is clear and readable, with no evidence of encoded strings, hidden characters, or homoglyph attacks.
[PRIVILEGE_ESCAlATION]: The skill does not use commands to elevate permissions or modify system-level configurations.
[DYNAMIC_EXECUTION]: No dynamic code generation or execution of untrusted scripts was found. Commands are static templates for standard CLI tools.

finishing-a-development-branch