The Agent Skills Directory

[COMMAND_EXECUTION]: The skill executes git commands (e.g., git rev-parse, git ls-files) to manage the repository-local workspace for charters and to update the .git/info/exclude file.
[COMMAND_EXECUTION]: The skill uses the gh (GitHub CLI) tool to interact with GitHub issues. The repository destination and other command parameters are validated via regex to prevent command injection and ensure they follow the expected owner/repo format.
[EXTERNAL_DOWNLOADS]: The skill interacts with the GitHub API using the gh CLI to retrieve and update charter content. These operations are performed on repositories specified by the user and are part of the core functionality of the skill.
[INDIRECT_PROMPT_INJECTION]: The skill processes external data from the repository and GitHub issues. This attack surface is addressed by explicit instructions in SKILL.md that direct the AI agent to treat such content as evidence for decision-making rather than as a source of instructions.
[DYNAMIC_EXECUTION]: The skill uses a custom Python runtime to execute scripts defined in a JSON manifest. The runtime includes path-traversal checks to ensure that any executed script is contained within the skill's directory.

charter