The Agent Skills Directory

[COMMAND_EXECUTION]: The skill is centered around the execution of gh CLI commands to interact with GitHub services. This is the primary and intended function of the skill, and commands are used in a standard, documented manner.
[INDIRECT_PROMPT_INJECTION]: The skill defines a surface for processing data from an external, potentially untrusted source (GitHub).
Ingestion points: Data enters the system via commands such as gh issue list, gh pr view, gh run view --log, and gh search.
Boundary markers: The provided examples do not include explicit instructions or delimiters for the agent to treat external content as untrusted.
Capability inventory: The skill enables various write operations including creating/editing issues and PRs, triggering CI workflows, and deleting repositories.
Sanitization: The skill promotes the use of --json and --jq for parsing, which provides structural safety but does not filter the text content for malicious prompt instructions. Given the intended use as a CLI reference, this is documented as an inherent surface.
[SAFE]: Documentation for installing the GitHub CLI using sudo apt install and brew install represents standard setup procedures for the required tool and is not a malicious privilege escalation attempt. All network operations are directed at official GitHub endpoints via the official CLI tool.

github