SmartBear plugin for Pact consumer tests, provider verification and Drift OpenAPI checks, with agents that write and review pact tests.
How it works
Foe: Bug Slime, a defect in your code.
Foes
- Bug SlimeHaze Poison Ward Insight Flicker Pinned Enraged
Party · Context
- SentinelYour agent100/100Haze Poison Ward Insight Flicker Pinned Enraged
- Pact generatorSubagent0/40Haze Poison Ward Insight Flicker Pinned Enraged
- Pact reviewerSubagent0/40Haze Poison Ward Insight Flicker Pinned Enraged
The Bug Slime blocks the way.
What really happensIt stands for a defect in your code.
Sentinel: Context 100/100.
The Bug Slime uses Regress.
What really happensA fix in one place breaks something that used to work.
Sentinel takes 8. Sentinel: Context 92/100.
The hero equips the PactFlow relic when two services must agree.
What really happensThe pactflow skill loads when you mention Pact, contract tests, provider verification, provider states or can-i-deploy.
Bug Slime takes 18.
An ally writes the first contract from the client's own code.
What really happensThe pact-generator agent writes Pact consumer tests and provider state handlers from client code, an OpenAPI spec or example requests.
Pact generator: Context 40/40.
The Bug Slime lashes at Pact generator.
What really happensThe subagent takes the hit in its own context. The main context stays clean.
Pact generator takes 8. Pact generator: Context 32/40.
The summons work the problem, each in its own context.
What really happensEach subagent does its own model and tool work, apart from the main conversation.
Bug Slime takes 8.
Pact generator returns with a short report.
What really happensOnly the subagent's final summary comes back to the main agent.
Sentinel takes 2. Sentinel: Context 90/100. Sentinel: Insight. Pact generator returns.
A second ally inspects every contract for false comfort.
What really happenspact-reviewer ranks issues by severity: exact matchers where types would do, random data, provider logic in consumer tests, missing states.
Pact reviewer: Context 40/40.
The summons work the problem, each in its own context.
What really happensEach subagent does its own model and tool work, apart from the main conversation.
Bug Slime takes 8.
Pact reviewer returns with a short report.
What really happensOnly the subagent's final summary comes back to the main agent.
Sentinel takes 2. Sentinel: Context 88/100. Pact reviewer returns.
Drift checks that the provider still keeps its word.
What really happensThe drift-testing skill writes and runs Drift CLI cases that verify a running API against its OpenAPI spec.
Bug Slime takes 36.
With a broker linked, the hero learns which builds may ship.
What really happensWith PactFlow credentials, the bundled SmartBear MCP server publishes contracts and runs can-i-deploy before a release.
Bug Slime takes 30. Bug Slime falls.
The Bug Slime falls.
Learn this spell
Claude Code
claude plugin marketplace add anthropics/claude-plugins-community
claude plugin install swagger-contract-testing@claude-communityCodex
npx skills add pactflow/pactflow-agent-skillsNotes
Writing and reviewing local Pact tests works without credentials. Publishing, can-i-deploy and the AI generate and review tools go through the bundled SmartBear MCP server and need a PactFlow account or Pact Broker plus Node.js 20. Drift tests need the Drift CLI; pact-coverage and the spec generators need ripwire. The plugin ships hooks and an MCP config, so trials need --allow-hazards, and an offline arena needs the Pact library already installed. Codex: npx skills add pactflow/pactflow-agent-skills adds the skills without the agents.
This spell has no recorded sessions in the curator's logs. Tried it? Send a spell report.