The Agent Skills Directory

[Indirect Prompt Injection] (LOW): The skill enables the agent to read content from GitLab issues, merge requests, and pipeline logs which can contain untrusted data.
Ingestion points: Commands like glab mr view, glab issue list, and glab ci trace ingest external data into the agent's context.
Boundary markers: None specified in the instructions to separate untrusted GitLab content from agent instructions.
Capability inventory: The skill allows the use of Bash which can be used to execute arbitrary commands if the agent is manipulated by content within a merge request or issue.
Sanitization: No specific sanitization or validation of the ingested content is mentioned.
[Credential Handling] (LOW): The skill provides instructions for managing GITLAB_TOKEN environment variables. While no secrets are hardcoded, users should be aware that the agent will have access to these credentials in the shell environment to perform its tasks.
[Command Execution] (SAFE): The skill requires the Bash tool to execute glab commands. This is consistent with the primary purpose of a CLI-focused skill. The instructions follow best practices for using the official GitLab CLI tool.

glab