The Agent Skills Directory

[SAFE]: No malicious patterns or security vulnerabilities were identified. The skill acts as a reference for standard functionality of the official GitLab CLI tool.
[COMMAND_EXECUTION]: The skill documents the use of the glab CLI for infrastructure-as-code tasks. It includes examples for initializing states and managing backends, including the use of flags to specify directories or alternative binaries.
[INDIRECT_PROMPT_INJECTION]: The skill provides an interface for the agent to ingest external data from GitLab infrastructure states, which represents a potential attack surface if the remote state contains untrusted content.
Ingestion points: Data is retrieved via glab opentofu state download and glab opentofu state list --output json from remote GitLab repositories.
Boundary markers: None are present in the provided documentation to distinguish between instructions and data.
Capability inventory: The agent can execute glab commands to modify infrastructure state (init, delete, lock, unlock).
Sanitization: No sanitization or validation of the retrieved JSON state data is described.

glab-opentofu