The Agent Skills Directory

[SAFE]: The skill implements a structured workflow for development tasks, focusing on verification and documentation rather than risky autonomous actions.\n- [DATA_EXPOSURE]: The workflow includes a specific security checkpoint requiring the agent to 'Review changed files for unintended edits and secrets,' which is a recommended practice to prevent accidental credential leakage.\n- [COMMAND_EXECUTION]: The skill utilizes local development commands (e.g., pnpm build, pnpm test) found within the project environment. These are restricted to the project's context and are standard for continuous integration/deployment tasks.\n- [PROMPT_INJECTION]: The skill ingests data from local project files such as TASKS.md and documentation files. While this represents a standard indirect injection surface, the skill contains no patterns suggesting susceptibility to malicious instruction override, and its capabilities are limited to standard git and build operations.

ship