foundry-quality

Plugin foundry-quality from marketplace foundry in fedcal/foundry. Commands read from the repo's manifest — 1 month ago.

Skills (5)

• define-slo

  Turn one user journey into SLIs, an SLO with a justified target, an error-budget policy with real consequences, and multi-window multi-burn-rate alerts. Use when reliability is discussed without numbers, before promising an availability figure, when alerts are noisy or ignored, or when an SLA is bei

• perf-budget

  Set, wire and enforce performance budgets in CI with real measured numbers and a gate that actually fails the build. Use before a launch, after a latency regression, when performance work keeps being given back, or when someone asks "is this fast enough" without a number to answer it. Produces budge

• postmortem

  Run a blameless postmortem that produces a timestamped timeline, multiple contributing factors, and actions each with a named person and a due date — then writes a Foundry runbook so the second occurrence is faster than the first. Use after any SEV1, any SEV2 that ran long, any incident consuming mo

• quarantine-flaky

  Detect, quantify, quarantine and actually fix flaky tests. Use when CI is red for no reason, when people re-run builds to get them green, when retries are hiding instability, or before adding tests to a suite nobody trusts. Every quarantine entry gets a named owner, a 14-day deadline and one of exac

• test-plan

  Build a risk-based test plan as a plan.v1 artifact, tied line by line to requirement.v1 acceptance criteria. Use when starting a test suite, when a release needs a defensible test scope, when coverage is being argued about without data, or when the suite is large but bugs still escape. Produces the

Subagents (6)

• contract-tester

  Consumer-driven contract testing between services and between frontend and backend — writes consumer expectations, wires provider verification into the provider's CI, detects breaking changes before deploy, versions and tags contracts per environment, and states plainly the failure modes contract te

• e2e-engineer

  Builds end-to-end suites that stay green for real reasons — selects the small set of journeys worth the cost, makes test data deterministic and isolated, sets an explicit stubbing policy at the network boundary, replaces every sleep with a state-based wait, captures trace/video/screenshot/console/ne

• observability-engineer

  Instruments a system with OpenTelemetry so incidents are diagnosable — structured logs carrying trace and correlation ids, RED metrics for request-driven services and USE metrics for resources, span design with deliberate cardinality control, a sampling strategy that keeps the errors and the tail, a

• performance-engineer

  Backend and system performance engineering driven by measurement — sets latency/throughput/resource budgets and SLIs before touching code, profiles to attribute cost to a named frame or query before optimising, designs load, soak, spike and breakpoint tests with a stated workload model, builds a cap

• sre-planner

  Turns user journeys into SLIs, SLOs with justified targets and an error-budget policy with teeth; designs multi-window multi-burn-rate alerts instead of raw thresholds; measures toil as a percentage of engineering time and sets a reduction target; defines incident roles and severity levels; and runs

• test-strategist

  Decides what to test, at which level, and what to deliberately leave untested — risk-driven test strategy for a specific codebase. Chooses pyramid vs testing-trophy shape from the code's actual failure modes, sets the integration/unit boundary, treats coverage as a diagnostic rather than a target, s

Similar plugins