The Agent Skills Directory

[CREDENTIALS_UNSAFE]: The skill explicitly reads and decodes sensitive Kubernetes secrets containing API keys. Evidence: kubectl get secret instantly-secrets -o jsonpath='{.data.api-key}' | base64 -d in SKILL.md.
[COMMAND_EXECUTION]: The skill executes high-privilege commands to modify the production environment, including changing environment variables and restarting services. Evidence: kubectl set env deployment/instantly-integration ... and kubectl rollout restart deployment/instantly-integration in SKILL.md.
[COMMAND_EXECUTION]: The skill performs credential modification by creating and applying new Kubernetes secrets. Evidence: kubectl create secret ... | kubectl apply -f - in SKILL.md.
[COMMAND_EXECUTION]: The skill invokes a local shell script whose contents are not provided for verification. Evidence: ./scripts/instantly-debug-bundle.sh in SKILL.md.
[INDIRECT_PROMPT_INJECTION]: The skill ingests potentially untrusted data from service logs and external status pages while having high-privilege cluster access. Ingestion points: kubectl logs and curl -s https://status.instantly.com. Capability inventory: kubectl modification and secret access. Sanitization: No sanitization or boundary markers are present for the ingested data.

instantly-incident-runbook