The Agent Skills Directory

[CREDENTIALS_UNSAFE]: The skill instructs the agent to retrieve and decode Kubernetes secrets, exposing sensitive API keys in plain text to the session output.
Evidence: kubectl get secret posthog-secrets -o jsonpath='{.data.api-key}' | base64 -d in SKILL.md.
[COMMAND_EXECUTION]: The skill attempts to execute a shell script located at ./scripts/posthog-debug-bundle.sh which is not provided as part of the skill definition, representing an unverifiable command execution risk.
Evidence: ./scripts/posthog-debug-bundle.sh in SKILL.md.
[COMMAND_EXECUTION]: The skill performs high-privilege administrative actions on the production infrastructure, including modifying secrets, updating environment variables, and restarting deployments.
Evidence: kubectl apply, kubectl set env, and kubectl rollout restart commands in SKILL.md.
[PROMPT_INJECTION]: The skill ingests untrusted log data from production applications, creating an attack surface for indirect prompt injection if an attacker can control logged content.
Ingestion points: kubectl logs -l app=posthog-integration --since=5m in SKILL.md
Boundary markers: Absent
Capability inventory: kubectl, curl across SKILL.md
Sanitization: Absent

posthog-incident-runbook