The Agent Skills Directory

[COMMAND_EXECUTION]: The skill uses kubectl and curl to perform system diagnostics and incident mitigation, such as restarting deployments and checking health status. These operations are restricted to the relevant application context.
[DATA_EXFILTRATION]: Accesses the JUICEBOX_API_KEY environment variable to authenticate requests to the official service domain (api.juicebox.ai). This follows standard operational practices for secret management.
[SAFE]: Evaluation of the indirect prompt injection surface identifies that the skill ingests data from logs and status APIs to function. No malicious behavior was detected.
Ingestion points: kubectl logs and curl status endpoints (references/implementation-guide.md).
Boundary markers: Absent.
Capability inventory: kubectl set env and kubectl rollout restart (references/implementation-guide.md).
Sanitization: Absent.

juicebox-incident-runbook