The Agent Skills Directory

[SAFE]: The skill implements a robust orchestration framework with predefined state transitions and explicit user-in-the-loop requirements at critical decision points, specifically using [Stop] markers to prevent unauthorized autonomy.
[COMMAND_EXECUTION]: Utilizes bash commands exclusively for standard repository management tasks, such as git commits and diffs, which are appropriate and necessary for its stated purpose of software implementation management.
[DATA_EXFILTRATION]: No evidence of network communication, external data transfers, or unauthorized access to sensitive system files (e.g., credentials, SSH keys) was found.
[PROMPT_INJECTION]: Employs a mandatory [SYSTEM CONSTRAINT] suffix for all sub-agent prompts to ensure they remain focused on the orchestrator's defined rules, which helps maintain execution stability and prevent instruction drift.

recipe-implement