Changelog

---

v0.1.0 — 2026-03-30

New: Observability, Storage & Graph Native Plugins

Five new built-in integrations shipping with the OSS engine:

Neo4j — query and assert on graph data using Cypher

• neo4j.query, neo4j.assert

MinIO / S3 — object storage operations in flows

• minio.put, minio.get, minio.delete, minio.assert

OTel / Tempo — trace context propagation and span assertion

• otel.inject, otel.assert

Grafana Loki — log query and assertion with LogQL

• loki.query, loki.assert

Prometheus — PromQL queries and metric counter assertions

• prometheus.query, prometheus.assert

New: LGTM Local Stack

./infra.sh up now starts the full LGTM observability stack alongside existing infrastructure: OTel Collector, Grafana Tempo, Grafana Loki, Prometheus, and Grafana (with all three datasources pre-configured).

Demo Services

All four demo services (user, product, order, notification) now emit:

• Structured JSON logs via zap, exported to Loki via OTLP
• Prometheus metrics (http_requests_total, http_request_duration_seconds, http_requests_in_flight) on GET /metrics
• Traces to OTel Collector (previously went directly to the TestMesh API)

---

v0.1.11 — 2026-04-15

GitHub PR Write-Back

• Analysis results posted directly to pull requests as comments with severity-tagged findings
• Commit status checks (pending → success / failure) set automatically on PR events
• Auto-fix PRs: self-healing suggestions above a confidence threshold create fix PRs automatically
• PR write-back configuration per integration (comment_on_pr, set_commit_status, auto_pr_enabled, auto_pr_threshold)

Workspace-Scoped Integrations

• Integrations (AI, Git, Notification) can now be configured per workspace, overriding global defaults
• Per-agent LLM provider routing: assign different AI providers to different agents within a workspace
• New WorkspaceAIConfig model with default provider and agent-level overrides
• Provider resolution chain: agent override → workspace default → global default → env var

Semantic Search & Embeddings

• Vector embeddings for graph nodes, flows, and code changes using OpenAI text-embedding-3-small
• pgvector-backed storage with HNSW indexing for fast similarity search
• Async embedding pipeline with 10-worker pool and bounded queue
• Coverage, impact, diagnosis agents and diff analyzer enhanced with semantic search
• New search API endpoints for nodes, flows, and code

System Graph CLI

• testmesh graph scan — scan local directories or remote Git repos; discovers services, APIs, DB tables, queues, and their relationships
• testmesh graph status — show node/edge/coverage/conflict statistics
• testmesh graph services — list discovered services
• testmesh graph show <node-id> — inspect a node and its dependencies
• testmesh graph coverage — coverage percentage with uncovered node list
• testmesh graph search <query> — search across all graph nodes
• testmesh graph conflicts — list and review merge conflicts
• testmesh graph export --format json|dot|mermaid — export for visualization

Claude Code Plugin

• Install via Claude Code: /plugin install testmesh
• Includes 8 skills: generate-flow, run-flow, validate-flow, action-patterns, debugging-flows, flow-layers, mcp-tools, yaml-schema

Installation

• Windows installer: irm https://testmesh.io/install.ps1 | iex
• npm package: npm install -g @testmesh/cli (was incorrectly documented as testmesh)

Also in this release

• mcp_call action — call external MCP servers as steps in a flow, enabling AI tool use within tests
• testmesh chat — conversational flow creation; describe what you want to test in plain English
• Plugin system for custom action handlers — bring your own protocol or internal tool
• Advanced reporting: JUnit XML output, HTML reports, and Slack notifications on failure

---

v0.4.0 — 2025-12-01

CLI

• testmesh generate --spec openapi.yaml — AI-powered flow generation from an OpenAPI spec
• testmesh watch — TDD mode; re-runs flows automatically when YAML files change

Actions

• websocket action type added — connect, send messages, and assert on received frames
• Mock server gains recording and playback mode — record real traffic and replay it in tests

Infrastructure

• Kafka SASL/TLS authentication support — connect to Confluent Cloud, MSK, and other secured clusters
• Dashboard: execution timeline view with per-step timing

---

v0.3.0 — 2025-09-15

Actions

• mock_server_start / mock_server_stop / mock_server_verify — spin up an in-process HTTP mock server as part of your flow setup
• grpc_call action type — call gRPC services with proto reflection or inline message definitions
• redis_get / redis_set action types — read and write Redis keys as flow steps

CLI

• testmesh debug — interactive step-by-step mode; inspect request, response, and extracted variables at each step

Testing

• Contract testing assertions — assert that a response matches a JSON Schema or OpenAPI response definition

---

v0.2.0 — 2025-06-01

Actions

• database_query action — run SQL queries against PostgreSQL and assert on results
• kafka_producer action — publish messages to Kafka topics
• kafka_consumer action — consume and assert on messages from Kafka topics; supports from_beginning and timeout

Flows

• Variable extraction with JSONPath — output: block with $.body.field syntax
• Retry configuration per step — max_attempts, delay, backoff: exponential | linear | constant

CLI

• testmesh validate — validate a flow YAML file without executing it

---

v0.1.0 — 2025-03-01

Initial release.

Actions

• http_request — HTTP/HTTPS requests with method, headers, body, and timeout

Flows

• YAML flow format with flow: wrapper at root
• assert: block with expression-based assertions (status == 200, body.id != "")
• env: block for flow-level environment variables

CLI

• testmesh run — execute a flow from a YAML file
• Exit codes: 0 (all passed), 1 (flow error), 2 (assertion failure)

Deployment

• Docker Compose configuration for local development
• API server on port 5016, Next.js dashboard on port 3000