glab-cluster
Pass
Audited by Gen Agent Trust Hub on Oct 31, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill instructs the agent to process outputs from commands like
glab cluster agent listandglab cluster agent token list, which ingest untrusted data from external GitLab repositories. - Ingestion points: Output from CLI commands querying GitLab agents, tokens, and cluster configurations from potentially untrusted projects (SKILL.md, references/commands.md).
- Boundary markers: No specific delimiters or warnings are used to separate untrusted CLI output from agent instructions.
- Capability inventory: The skill allows the agent to interact with the GitLab CLI, which has the capability to modify cluster configurations, manage tokens, and update
kubeconfig(references/commands.md). - Sanitization: No evidence of sanitization or validation of the CLI output is provided in the instructions.
- [DATA_EXPOSURE_&_EXFILTRATION]: The skill documentation explicitly references sensitive files like
~/.kube/configand~/.kube/config.lock. While accessing these files is the primary purpose of the skill (managing Kubernetes clusters), agents should handle this data with high caution to prevent unintended exposure of credentials.
Audit Metadata