The Agent Skills Directory

[PRIVILEGE_ESCALATION]: The scripts/gh-update.sh script utilizes sudo commands to modify system-level directories and configuration files, including /etc/apt/keyrings and /etc/apt/sources.list.d. It performs administrative package management operations (sudo apt update and sudo apt install) which grant the skill elevated control over the host environment.
[EXTERNAL_DOWNLOADS]: The skill fetches an archive keyring from the official GitHub CLI domain (https://cli.github.com/packages/githubcli-archive-keyring.gpg). While this is a well-known and legitimate source, the process involves downloading and executing installation logic at the system level.
[COMMAND_EXECUTION]: The skill executes multiple shell commands through the GitHub CLI (gh), including creating issues with user-supplied titles/bodies and performing GraphQL API mutations.
[PROMPT_INJECTION]: The skill exhibits an indirect prompt injection surface by ingesting untrusted data.
Ingestion points: User-provided $ARGUMENTS and terminal error messages/tracebacks are used to populate issue templates.
Boundary markers: None. There are no delimiters or instructions used to separate the user-provided data from the agent's internal logic.
Capability inventory: The skill can execute shell commands (gh), perform network operations via API, and modify the system via the gh-update.sh script.
Sanitization: No evidence of sanitization, filtering, or escaping is present for the terminal logs or user input before they are interpolated into the GitHub issue body.

create-issue