generated: '2026-08-29' method: generated source: >- openapi/spyderbat-openapi.json (every operationId verified present in the spec) plus https://docs.spyderbat.com/ product documentation note: >- Packaged Agent Skills for the marquee Spyderbat flows. Every operationId referenced was grepped out of the saved OpenAPI before the skill was written; no operation, parameter or error is invented. Skills deliberately record what Spyderbat does NOT publish (idempotency, retry policy, webhook signing, reversal windows) because those absences are what an agent needs to know before it acts. skills: - name: spyderbat-api-key-and-org-setup file: spyderbat-api-key-and-org-setup.md summary: >- Create a key, find the org UID, understand the RBAC role it inherits, make the first call. The prerequisite for every other skill. operations: 7 risk: read-only - name: spyderbat-triage-spydertraces file: spyderbat-triage-spydertraces.md summary: >- Work the alert queue — search model_spydertrace by score and time window through the submit-then-poll job API, pull the trace contents, attribute the workload. operations: 7 risk: read-only - name: spyderbat-suppress-noisy-detections file: spyderbat-suppress-noisy-detections.md summary: >- Preview a suppression policy and its scope with the only dry-run affordance in the API, then apply it with human confirmation. operations: 5 risk: write — no documented unsuppress operation - name: spyderbat-agent-response-actions file: spyderbat-agent-response-actions.md summary: >- Kill a pod, kill a process, or scan a container on a live monitored host. The only rate-limited operations in the contract, and the only irreversible ones. operations: 6 risk: destructive — irreversible, no dry run, no idempotency key - name: spyderbat-forward-events-to-siem file: spyderbat-forward-events-to-siem.md summary: >- Enable SIEM forwarding on a saved query and consume the events iterator, or configure email/Slack/PagerDuty/webhook notification targets and test them. operations: 14 risk: write — configuration only