The Agent Skills Directory

[Indirect Prompt Injection] (LOW): The skill provides scripts and workflows that ingest untrusted data from external sources such as application logs and Prometheus metrics.
Ingestion points: references/incident-response.md (functions using kubectl logs and prometheus_api_client queries).
Boundary markers: Absent in example scripts.
Capability inventory: kubectl (rollout, exec, set image), git (push/clone), and gh (repo management) across multiple reference files.
Sanitization: No explicit sanitization of log content or metric labels before processing.
[Command Execution] (LOW): The skill contains references to high-privilege commands like kubectl exec and tcpdump for incident forensics in references/incident-response.md. These findings are downgraded to LOW as they are intrinsic to the primary DevOps/SRE purpose.
[External Downloads] (LOW): The skill references standard industry tools (git, gh, kubectl) and trusted GitHub Actions (actions/checkout, docker/setup-buildx-action). These align with the [TRUST-SCOPE-RULE] and are consistent with DevOps workflows.

devops-engineer