The Agent Skills Directory

[COMMAND_EXECUTION]: The skill documents the valueFromCommand setting within custom_headers, which allows the glab CLI to execute a specified external command to retrieve header values. This results in the execution of the command every time glab makes a request to the configured host (SKILL.md, references/commands.md).
[PRIVILEGE_ESCALATION]: The skill provides instructions for enabling duo_cli_auto_run and orbit_local_auto_run. These settings suppress the interactive confirmation prompts usually required when glab invokes the GitLab Duo CLI or Orbit local binaries, increasing the autonomy of automated processes (SKILL.md, references/commands.md).
[CREDENTIALS_UNSAFE]: The skill details the management of sensitive credentials, including GitLab personal access tokens, OAuth client secrets, and private keys. It emphasizes security best practices, such as using environment variables (valueFromEnv) or credential helpers instead of storing literal secrets in the YAML configuration file (SKILL.md, references/commands.md).
[COMMAND_EXECUTION]: The skill allows the configuration of global preferences for the system editor and glab_pager, which controls which external applications are launched by glab for editing and viewing output (SKILL.md, references/commands.md).

glab-config