What problem does it solve? Setting up a performance testing pipeline from scratch involves scaffolding k6 helpers, discovering endpoints, authoring scenarios, running baselines and load passes, and gating results against SLOs — a process that is easy to leave half-finished. This Skill orchestrates that entire lifecycle autonomously, with reviewer-gated phase transitions and hard safety ceilings so no step is skipped. ## Core Features & Use Cases - Seven-phase pipeline: Scaffold, Readiness, Scenario-model, Baseline, Load-run, Threshold-gate, and Report, each with explicit deliverables, exit criteria, and reviewer approval before advancing. - Derive-or-bootstrap readiness probe: Reuses journey maps and HAR captures from prior functional onboarding when available, or falls back to crawler/OpenAPI endpoint discovery. - Enforced safety model: Hard ceilings of 1000 VUs and 3600 seconds per run, origin allowlisting, and explicit opt-in required before any production load testing. - Threshold enforcement proof: A mandatory deliberate-breach test proves k6 thresholds actually fail builds before SLO verdicts are trusted. - Use Case: Point the Skill at a staging origin and receive a maintained perf suite with scenario scripts, baselines, load/stress/spike/soak results, SLO verdicts, and a stakeholder-readable report. ## Quick Start Ask the agent to perf-onboard your project against your staging URL, for example by saying "perf-onboard https://staging.example.com". ## Quick Start Invoke the skill with a target origin such as "perf-onboard https://staging.example.com" and the pipeline runs all seven phases autonomously.
