Works with the AI Agents you already use

    Claude CodeClaude CodeCursorCursorCodex CLICodex CLIDeepSeekDeepSeekMuseMuse+22 more

    .NET API Contract Guard

    1

    Review breaking ASP.NET Core API changes before release. Compare source or OpenAPI contracts, show the old request or client expectation at risk, and get focused compatibility fixes with clear evidence and limits.

    Free

    Security scanned
    .NET API Contract Guard

    .NET API Contract Guard

    Example session with this skill installed

    Compare before.json and after.json for existing-client compatibility. Review POST /orders. Show the old request or consumer expectation affected by each supported contract break. Do not change either file or infer runtime status codes from the documents.

    • Read your context and instructions
    • Compiled the net api contract

    Local demonstration: declared contract comparison

    Date: 2026-11-03. Original synthetic data, reviewed manually in the authoring chat using the skill's workflow. The JSON files were parsed locally to inspect input requirements and output fields. No server or client was executed. This is a worked example, not an independent behavioral benchmark.

    Baseline: before.json, declared API version 1.0. Candidate: after.json, declared API version 1.1. Scope: the sole operation POST /orders, its application/json request and its declared 200 application/json response. Both inputs use OpenAPI 3.0.3 and inline schemas; no references require resolution.

    Outcome: two supported changes to the declared contract affect existing consumers. Runtime enforcement remains unverified.

    The previously optional reference input is now required. Both documents retain sku and reference as string properties, but the candidate adds reference to the required array. Evidence in each document: /paths/~1orders/post/requestBody/content/application~1json/schema/required. The old payload {"sku":"DEMO-1"} satisfies the baseline schema and fails the candidate schema because reference is absent. This narrows accepted input in the declared contract. Preserve reference as optional with suitable behavior for omission, or introduce a versioned request contract if the requirement is intentional. Verify through an isolated host using the same old payload against both revisions; no rejection status code has been established here.

    The required response identifier changes from orderId to id. Evidence: before.json /paths/~1orders/post/responses/200/content/application~1json/schema/properties/orderId and its required array; after.json the corresponding properties/id and required array. An old consumer expecting orderId, previously guaranteed by the response schema, no longer has that guarantee. The candidate schema permits a response such as {"id":"order-demo"} without orderId. Preserve the original wire name or return a compatible old-version response. Verify the actual emitted response and a representative old consumer before claiming an observed runtime failure.

    No route, HTTP method, media type or success status change appears in these inputs. Changing info.version does not introduce a separate versioned route or preserve the old input contract by itself.

    Coverage: both complete supplied documents were inspected; one operation, one request schema and one response schema. Source code, serializers, validators, deployed endpoints, generated SDKs and actual consumer implementations were not supplied. No runtime test was run, and no actual customer impact or usage volume is asserted. A compatible rollout must resolve the new input requirement and retain the old output guarantee or deliberately migrate the affected consumers.

    before.json

    Connects securely to your tools. The creator never sees your data.

    About this skill

    An API can compile successfully while a previously valid client request stops working.

    .NET API Contract Guard guides your coding agent through a focused backward-compatibility review. Give it a PR or source comparison, or two OpenAPI documents. It returns an evidence-backed report linking each supported break to the affected operation and an existing client request or response expectation.

    What you get

    • Reviews of routes, parameter binding, required inputs, JSON field names, response types and status changes.
    • Separate treatment of request and response compatibility, including shared DTOs.
    • A distinction between supported breaks, conditional client risks and incomplete coverage.
    • Concrete options such as preserving a JSON alias, keeping an input optional or retaining an old route.
    • Focused regression-test guidance when tests are requested.

    How to use it

    Ask: "Review this ASP.NET Core PR for breaking API changes against the release baseline." Name the baseline and candidate, or attach the two OpenAPI documents. Known client requirements make the conclusion more precise.

    The skill reviews first. A review request does not silently become an implementation task. If you ask for fixes or tests, it follows the repository's conventions and reports what actually ran.

    Where it stops

    This is an agent instruction skill, not a standalone scanner, CI integration or compatibility guarantee. It cannot establish runtime behavior from a schema alone. It does not include NuGet binary compatibility, EF migrations, penetration testing or hosted API monitoring. With no baseline, it reports current-state risks rather than claiming a proven regression.

    Find more skills like this on the Claude Skills Marketplace, the Codex Skills Marketplace and the Gemini Skills Marketplace, or browse Claude Skills for APIs & Backend.

    Agent instruction workflow, not a standalone scanner or CI gate. Document-only review cannot establish runtime behavior. No baseline means current-state risks only. Excludes NuGet binary compatibility, database migrations and penetration testing. No independent or cross-agent benchmark has been run. Cover artwork is an AI-generated illustration, not an execution screenshot.

    How to use

    1. 1

      Connect your agent to the Agensi MCP

      Copy the prompt and paste it into Claude, ChatGPT, Cursor or any agent that supports MCP. It sets up the connection for you. Manual setup

    2. 2

      Sign in to Agensi

      Your agent shows a sign-in link. Sign in and click Allow.

    3. 3

      Ask for what you need

      Your agent loads this skill and uses it on your task. It always asks before buying a skill.

    Reviews

    No reviews yet

    Be one of the first to try it. Every listed skill goes through the trust checks below.

    Security scanned

    Checked by our 8-point scan before listing

    Fresh listing

    Recently published to Agensi

    Free to use

    No account required to browse

    Trust & safety

    Security scan

    Security scan passed 14 days ago

    • Free with an Agensi account

    Software
    Git
    Optional
    .NET SDKProject test dependencies

    Git compares source revisions. The .NET SDK and test dependencies are only for runtime verification.

    Permissions

    Read Files
    Terminal / Shell
    Write Files
    Browser
    Network Access

    Review reads supplied source/OpenAPI files. Requested fixes or tests may write files and run shell commands. Official documentation may be consulted for version-sensitive questions. No connected account, API key or external comparison service is required; repository contents are not uploaded to comparison services.

    Listed14 days ago

    What's inside

    Frequently Asked Questions

    No. It gives your agent a focused review workflow and report format. There is no bundled executable or automatic CI gate.

    No. Source comparisons are supported. Two OpenAPI files also work when source is unavailable, with conclusions limited to the declared contract.

    No. Request and response direction matters. Strict-consumer risks are conditional unless actual consumer evidence supports a break.

    When requested, using the project tools and isolated test host. Test execution depends on available dependencies and repository configuration.

    No. The report states what was reviewed, what was inferred, what was executed and what remains unknown.

    New in APIs & Backend

    Free