gitops-status
Pass
Audited by Gen Agent Trust Hub on Apr 1, 2026
Risk Level: SAFEDATA_EXFILTRATIONCOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
- [DATA_EXFILTRATION]: The skill accesses operational data by fetching dashboard summaries, resource trees, and raw pod logs via the harness_get tool.- [COMMAND_EXECUTION]: The skill instructs the agent to perform administrative actions, specifically triggering GitOps application synchronization using the harness_execute tool.- [PROMPT_INJECTION]: An indirect prompt injection surface is present because the skill ingests external data from pod logs and application events.
- Ingestion points: Data enters the context via harness_get (resource_type: gitops_pod_log) and harness_list (resource_type: gitops_app_event).
- Boundary markers: None are present to distinguish logs or event data from system instructions.
- Capability inventory: The agent can execute sync actions via harness_execute and read further data via harness_get.
- Sanitization: There is no evidence of sanitization or filtering applied to the retrieved log or event content.
Audit Metadata