The Agent Skills Directory

[COMMAND_EXECUTION]: The skill instructs the agent to execute various glab orbit commands to interact with GitLab's Knowledge Graph service, including health checks and remote query execution.
[EXTERNAL_DOWNLOADS]: The glab orbit setup command is used to download and install a local Orbit CLI binary. This is a documented experimental feature of the official GitLab CLI (glab) and targets well-known GitLab infrastructure.
[PROMPT_INJECTION]: The skill is susceptible to indirect prompt injection because it retrieves and processes content from the GitLab Knowledge Graph, which contains data from untrusted sources (e.g., merge requests, issues).
Ingestion points: Untrusted data enters the agent context through the output of glab orbit remote query, glab orbit remote schema, and glab orbit remote graph-status.
Boundary markers: Absent. The skill instructions do not provide delimiters or warnings to the agent to disregard instructions embedded within the retrieved graph data.
Capability inventory: The agent has the ability to execute CLI commands (glab), perform network requests (via glab), and read local files (to load query envelopes).
Sanitization: Absent. There is no evidence of filtering, escaping, or validation of the graph data before it is presented to the agent.

glab-orbit