MCP Release Kit — Contract Regression & Failure Lab

Catch tool-contract drift before your next MCP release.

MCP Release Kit gives developers building MCP servers for AI clients a ready-made regression lab: source code, worked failure scenarios, expected detections and review guides. €29 one-time, including applicable tax.

INCLUDED

• 33 synthetic scenarios: 14 contract-change fixtures and 19 stdio healthy/failure modes.

• A strict baseline gate for tool presence, descriptions, input/output schemas and behavior hints.

• A deliberately broken demo server covering timeouts, malformed JSON, wrong response IDs, pagination loops, tool errors and changed outputs.

• Explicit tool-call cases with JSON-pointer assertions and a mandatory tool allow-list.

• HTML, JSON and aggregate JUnit reports; a GitHub Actions recipe; sample reports; integration, troubleshooting and release-review guides; AI-assisted review prompts.

• Editable Python source and 22 automated tests. Python 3.10+, standard library only. No LLM API calls or model credits needed for deterministic tests.

Start from the assembled fixtures and documented workflow, then adapt the baseline and assertions to your own server. Run the complete synthetic lab with one command. Download the free preview for sample reports, quickstart and compatibility details.

COMPATIBILITY

Live testing supports stdio MCP 2025-11-25 only. Offline comparison accepts complete exported tools/list catalogs. No live HTTP, OAuth or newer 2026 protocol support. Tested locally against the included synthetic lab; your server needs its own acceptance test.

This is a conservative contract-review gate, not full JSON Schema validation, a security scanner, MCP conformance certification or AI-agent evaluation. Free MCP Inspector is an alternative for general inspection. No hosted execution, managed support, SLA or future-update commitment is included.

License: internal use and modification, including internal commercial projects; no redistribution or resale.

Seller: Balaji. Support: balajimurugesan.2016@gmail.com and preethisivakumar94@gmail.com.