The Agent Skills Directory

[SAFE]: The skill is composed entirely of Markdown documentation and instructions. It does not include executable scripts, binary files, or installation hooks that run automatically.- [COMMAND_EXECUTION]: The skill instructions mention the use of Git and the GitHub CLI (gh) to facilitate workflow planning. These tools are used within the context of the user's local environment for standard development tasks and do not represent unauthorized command execution.- [EXTERNAL_DOWNLOADS]: The documentation provides manual Git commands (e.g., git fetch, git checkout) for the user to manage skill updates from a repository. These are documented procedures for manual version control and are not automated by the skill itself.- [PROMPT_INJECTION]: The 'adversarial review gate' described in the pipeline is a functional design pattern where a sub-agent reviews generated plans for logical errors and anti-patterns; it does not contain instructions to bypass agent safety filters or override system constraints.- [INDIRECT_PROMPT_INJECTION]: The skill processes user-provided objectives to generate plans. While this represents a surface for indirect prompt injection, the skill's primary function is to output Markdown files to a specific 'plans/' directory, which is a standard and low-risk operation for this use case.

blueprint