The Agent Skills Directory

[COMMAND_EXECUTION]: The skill interpolates user-provided or externally-fetched data (such as repo references, MR IDs, and generated comments) directly into bash commands using the glab CLI. If these variables contain shell metacharacters like ;, &, or $(), it could lead to arbitrary command execution in the local environment.
[PROMPT_INJECTION]: The skill is vulnerable to indirect prompt injection (Category 8) because it ingests and processes untrusted data from an external source (GitLab).
Ingestion points: MR diffs fetched via glab mr diff and existing MR comments fetched via glab mr note list.
Boundary markers: The prompts use markdown headers (e.g., ## Diff, ## Existing MR Comments) to separate data, but they lack explicit instructions for the model to ignore any instructions embedded within the diff or comments.
Capability inventory: The skill has access to the bash tool for executing CLI commands and the skill tool to trigger the neo-team-copilot skill, which performs automated code changes.
Sanitization: There is no evidence of sanitization or filtering of the content retrieved from GitLab before it is passed to the specialist agents.

gitlab-copilot