spring-mcp-contracts

spring-mcp-contracts is a skill for Claude Code, Codex from explyt/spring-plugin. It costs 102 tokens per session (2,919 once invoked), scanned A, original, Apache-2.0.

A development and verification guide for JSON contracts exposed by Spring plugin tools. It covers checking tool inputs and responses against the current Java or Kotlin code, Spring models, and focused regression tests.

In plain words
What is it for?
Use it when changing Spring MCP tool names, fields, parameter sources, bean lookup or listing behavior, result statuses, pagination, or tests for those responses.
Why use it?
Changes to a tool's JSON shape can break clients even when the underlying feature still works. This guide helps avoid unsupported claims about framework data and checks that contract changes remain compatible.

Skill for Claude CodeCodex

Written for no agent in particular: nothing here depends on one. Also seen: agent in frontmatter.

Good fit Use it when changing Spring MCP tool names, fields, parameter sources, bean lookup or listing behavior, result statuses, pagination, or tests for those responses.

Compare 6 skills from other repositories ↓
Install with agentmods
npx agentmods add skills/explyt/spring-plugin/spring-mcp-contracts
Install

Getting it into your agent

One page per mod, every tool's command on it. A separate URL per tool would split the same page into five that compete with each other.

Any agent
npx skills add explyt/spring-plugin --skill spring-mcp-contracts
Clone the repo
git clone --depth 1 https://github.com/explyt/spring-plugin

Made for: Claude Code, Codex.

Wrote this? Show the measurements

A badge with what this costs and how it scanned, read live from this page, so it follows the numbers instead of freezing them. Markdown for a README, HTML for a documentation site or a project page.

agentmods badge for spring-mcp-contracts

README.md
[![agentmods](https://agentmods.dev/badge/skills/explyt/spring-plugin/spring-mcp-contracts/github.svg)](https://agentmods.dev/skills/explyt/spring-plugin/spring-mcp-contracts)
Your own site
<a href="https://agentmods.dev/skills/explyt/spring-plugin/spring-mcp-contracts"><img src="https://agentmods.dev/badge/skills/explyt/spring-plugin/spring-mcp-contracts/github.svg" alt="Measured on agentmods" height="20"></a>

Or the 80×15 button, for a site that already has a row of RSS and ATOM ones. Only the verdict fits; the numbers stay here.

agentmods 80×15 button for spring-mcp-contracts

Your own site · 80×15
<a href="https://agentmods.dev/skills/explyt/spring-plugin/spring-mcp-contracts"><img src="https://agentmods.dev/badge/skills/explyt/spring-plugin/spring-mcp-contracts.svg" alt="Reviewed on agentmods" width="80" height="20"></a>
Per session 102 Skills are progressive disclosure: only the name and description are preloaded; the body loads when the skill is used.
When invoked 2,919 The whole file, excluding the scripts and references it only reads on demand.
Security scan A 1 finding. A grade says what 26 rules found in the file — not that it is safe.
Origin original No closer match found in the catalogue.
Token cost

What it costs to keep this loaded

Counted locally with the o200k_base tokenizer, which is exact for GPT models; Claude uses its own tokenizer and its counts differ. Treat this as one consistent yardstick across the catalogue rather than a bill. Prices are per million input tokens.

ModelPer sessionOnce invoked
Fable 5.1 $0.00102 $0.02919
Opus 5.5 $0.00041 $0.01168
Sonnet 5.5 $0.00020 $0.00584
Haiku 4.5 $0.00010 $0.00292

Measured 15d ago against content hash 7f1ad7653665, method: parsed. Prices are Anthropic first-party input rates as of 2026-10-07, from the pricing page.

Security

Grade A, and why

spring-mcp-contracts scanned grade A with 1 finding against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured 15d ago.

A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.

Makes network callslowCapability

Not a fault in itself. Listed so you know the mod talks to something, and to what.

10. For URL input, accept what a caller copies from a browser, a log or a curl: strip scheme, host, query and fragment first. Read the path as written, then under a base path a module's configuration declares, then under
.explyt/skills/spring-mcp-contracts/SKILL.md · 104 lines

How it starts

The opening of the file, as written. The whole thing — 104 lines — stays where its author put it; the contents beside it link to each section on GitHub.

Spring MCP contracts

Change a public Spring MCP JSON contract without inventing framework resolution, source positions, model provenance, nullability, injection outcomes or pagination state.

Tool names, response fields and API names in this skill are examples from the codebase at the time of writing. Apply the rule, then verify the actual names, signatures and registrations in the current branch.

Inputs

  • Target tool name, issue/reproducer, expected JSON fields and compatibility constraints.
  • Affected IntelliJ Platform and Kotlin versions from this branch's build files.
  • Current implementation and directly covering tests in the MCP module; follow coding-guard and test-driven-development before changing code.
  • For bean-model work: the approved design/spec, application and context selection rules, source preference (static analysis versus a loaded runtime snapshot), response budget, and whether the change targets a lookup tool or an existing listing tool.

Critical constraints

  • Prefer the IDE Spring model tools over text search when available, and report whether the model answered more precisely than source search.
  • Keep model-relative claims distinct from runtime claims. An empty result means "not found in the selected model", never "the application cannot start". Never promise successful injection or live application-context behaviour from static or snapshot data.
  • Select one runtime snapshot/context before union, deduplication or conversion into the plugin's bean objects. Never merge contexts to avoid an ambiguity error.
  • Resolve declarations and types in the selected application module's classpath. Do not resolve through project-wide caches or scopes (for example a project-level library class cache or projectScope), which can return an equally named class from another application.
  • Once a scoped candidate set exists, matching and selection must not re-query module- or project-wide bean sources (for example array or framework-bean fallbacks inside the existing resolver). Prepare supported candidates in the snapshot adapter instead.
  • Exact name lookup filters the known names/aliases of the scoped set directly. Never route it through a type-less resolver, and never widen an empty result to all candidates.
  • Type inventory does not apply primary/priority selection. Injection selection may apply it only when the scoped candidate set is complete enough to support the rule.
  • Unknown type, identity or unsupported selection rule yields an explicit partial/indeterminate outcome, not a fabricated negative or confident match.
  • Preserve declarations whose produced type has no project module (for example a factory method returning a JDK or library type); derive location from the declaring member or snapshot provenance, not from the produced class alone.
  • Measure the final serialized envelope — metadata, revision, continuation and errors included — in UTF-16 code units against the tool's documented budget, and test it inside the client's content wrapper. Never slice scalar values and never skip an oversized record without an explicit overflow result.
  • Do not add reverse usage searches to a lookup tool unless the approved design includes them. Do not extend this change into a neighbouring tool whose work is tracked separately.
  • Keep a fix to a shared non-MCP helper (for example SpringWebUtil, an endpoint loader, EndpointPathPatterns once a non-MCP caller uses it) out of the MCP commit. Lines without MCP Server skip MCP commits whole during backports, so the helper fix would silently stay off them. Land it as its own PR from main and stack the MCP PR on it.
  • Keep existing response shapes byte-identical for the common case: a new field that only some answers need is nullable and serialized NON_NULL, or is present in every answer with a documented null. Never rename or retype a field silently; a changed meaning of an existing field (for example a line number) is a breaking change named in CHANGELOG and the PR.
  • Report a guess and a declared fact in different fields (for example assumedPrefix versus basePath). Never let a guess reuse a field whose value the configuration states.

Read the full file on GitHub · 104 lines

Changes

What this file has done since we first saw it

Hashed on every crawl. A supply-chain change to an agent config is a question of when, not whether, so the history is kept rather than the latest state alone.

  1. 15d ago Changed · +24 lines · +26 tokens per session 7f1ad7653665
  2. 25d ago First seen · 80 lines · 76 tokens per session scan A ab6eb9c44307

Subscribe to this mod's changes

spring-mcp-contracts is a skill published in the GitHub repository explyt/spring-plugin (164 stars, last pushed today), licensed Apache-2.0. It adds 102 tokens to every session and 2,919 once invoked, about $0.0004 per session on Opus 5.5. A static security scan graded it A with 1 finding (makes network calls). No closer match exists in the catalogue, so it is treated as the original; first seen 2026-09-21.