More screenshots

    Works with the AI Agents you already use

    Claude CodeClaude CodeCursorCursorCodex CLICodex CLIMuseMuseDotsDots+21 more

    🛡️OpenAPI Contract Guard

    1

    Use it before accepting an API change, then check the findings against actual consumer expectations.

    $12

    Secure checkout via Stripe

    Money-back guarantee

    Converts to your local currency at checkout

    Security scanned
    🛡️OpenAPI Contract Guard

    🛡️OpenAPI Contract Guard

    Example session with this skill installed

    Using the bundled synthetic examples, run python scripts/scan_openapi_contract.py fixtures/risky-sample --format markdown from the package folder. Return the complete report and explain its limits. Do not modify any specification.

    • Read your context and instructions
    • Compiled the openapi contract guard

    Representative synthetic run

    This is the complete scanner output from package v1.0.1 against its bundled risky fixture, not a customer result.

    OpenAPI Contract Guard Report

    Files scanned: 3
    Findings: 13

    Coverage complete for supported heuristic scope: True
    Diff performed: True

    [MEDIUM] OAS002 - fixtures\risky-sample\openapi-v1.json:56

    Success response is missing a schema.
    Remediation: Declare a response schema.

    [MEDIUM] OAS001 - fixtures\risky-sample\openapi.json:12

    Operation is missing operationId.
    Remediation: Add a stable operationId.

    [MEDIUM] OAS003 - fixtures\risky-sample\openapi.json:16

    Parameter is missing a type.
    Remediation: Add a schema type for the parameter.

    [HIGH] OAS004 - fixtures\risky-sample\openapi.json:12

    Non-public operation lacks an explicit security requirement.
    Remediation: Add operation or global security, or mark the operation public.

    [LOW] OAS005 - fixtures\risky-sample\openapi.json:12

    Array schema is unbounded.
    Remediation: Add maxItems where payload size matters.

    [LOW] OAS005 - fixtures\risky-sample\openapi.json:12

    String schema is unbounded.
    Remediation: Add maxLength where payload size matters.

    [CRITICAL] OAS102 - fixtures\risky-sample\openapi-v2.json:4

    Operation was removed.
    Remediation: Restore the operation or version the API change.

    [CRITICAL] OAS101 - fixtures\risky-sample\openapi-v2.json:1

    Path was removed.
    Remediation: Restore the path or version the API change.

    [HIGH] OAS103 - fixtures\risky-sample\openapi-v2.json:4

    Response code was removed.
    Remediation: Keep documented response codes or version the API change.

    [CRITICAL] OAS104 - fixtures\risky-sample\openapi-v2.json:9

    Parameter became required or a required parameter was added.
    Remediation: Keep the parameter optional or version the API change.

    [HIGH] OAS106 - fixtures\risky-sample\openapi-v2.json:9

    Parameter enum was introduced or narrowed.
    Remediation: Keep accepted values or version the API change.

    [HIGH] OAS105 - fixtures\risky-sample\openapi-v2.json:19

    Parameter type changed.
    Remediation: Avoid narrowing or changing parameter types in-place.

    [HIGH] OAS107 - fixtures\risky-sample\openapi-v2.json:9

    Response schema field was removed.
    Remediation: Preserve response fields or version the API change.

    How to use this result

    Review each finding against the API and its consumers. Coverage is complete only for the supported heuristic scope, not full OpenAPI conformance. A comparison ran, but that does not prove compatibility. The scanner did not modify the inputs. Decide separately whether to restore, version or otherwise address each change.

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

    About this skill

    Review API changes with evidence

    OpenAPI Contract Guard reads local OpenAPI 3.0/3.1 JSON specifications and produces a reviewable lint and change report. Use it before accepting an API change, then check the findings against actual consumer expectations. It is a bounded heuristic, not a guarantee of backward compatibility.

    Inputs and outputs

    Provide one JSON specification for linting, or an unambiguous old/v1 and new/v2 pair for comparison. Receive JSON or Markdown findings with rule, severity, file, approximate line, remediation, coverage status and whether comparison ran.

    What it checks

    • Missing operation IDs, schema types, selected security requirements and unbounded strings/arrays.
    • Removed routes, operations, response codes and top-level response fields.
    • Required/type/enum changes for schema-based and single-media content-based parameters, including path inheritance and operation overrides.
    • Parameter representation or media-type changes requiring serialization review.
    • Ambiguous normalized media keys and malformed inputs, without silently treating skipped checks as a clean result.

    Requirements and boundaries

    An agent must read the complete package and execute Python 3 with local file access. The scanner uses the standard library only and does not fetch references, contact services or rewrite specifications. Convert YAML to JSON separately. Swagger 2 and reference-containing documents are excluded. Nested schema semantics, request bodies, composition and actual wire compatibility need separate review. An exit code of zero means a report was produced, not that compatibility passed.

    Package update

    Version 1.0.1 repairs the original scanner and instructions, improves malformed-input handling and parameter/media comparisons, and aligns the documentation. Synthetic local and independent Claude execution tests passed; this is execution evidence, not buyer demand or proof of compatibility. Existing price and licence terms are unchanged.

    Heuristic only; not full conformance or backward-compatibility proof. No native YAML, Swagger 2, reference resolution, nested response-property comparison, request-body compatibility or composition analysis. BOM-prefixed JSON, boolean parameter schemas and RFC 2231 media parameters are outside supported scope. ASCII control characters in media keys, ambiguous normalized media keys and out-of-range numbers produce incomplete coverage. Literal $ref keys are conservatively excluded. Inspect findings, coverage_complete and diff_performed; exit zero is informational.

    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 any payment.

    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

    Money-back guarantee

    Doesn't do what its page promises? You get your money back.

    Trust & safety

    Security scan

    Security scan passed 20 days ago

    Requires

    Python 3

    Permissions

    Terminal / Shell
    Read Files

    File Scopes

    **/openapi*.json
    **/openapi*.yaml
    **/openapi*.yml
    **/swagger*.json
    **/swagger*.yaml
    **/swagger*.yml
    **/*.openapi.json
    **/*.openapi.yaml

    Read-only local input access and Python execution. No network, environment-variable access or specification writes. Explicit files can have any name; directory discovery is restricted. YAML conversion is external to the scanner. Unsupported inputs produce incomplete coverage, not a compatibility pass.

    Listed5 months ago
    Updated20 days ago

    What's inside

    Frequently Asked Questions

    No. It surfaces supported contract risks for review. Confirm consumer expectations and run appropriate API tests before accepting a change. No findings is not proof of compatibility.

    The scanner reads OpenAPI 3.0/3.1 JSON. Convert YAML separately before scanning. Swagger 2 and documents containing references are explicitly excluded; conversion of syntax does not convert the API version or resolve references.

    Required, type and enum changes, including array-item enums, for schema-based and single-media content-based parameters. Path parameters are merged with operation overrides. Representation or media changes are flagged for manual serialization review.

    SKILL.md, Python scanner, README, agent metadata, review references, synthetic fixtures and bundled tests. Use the complete package. Existing licence terms remain unchanged.

    Local file access, Python 3 and permission to execute the scanner. It uses the standard library only. Independent Claude synthetic execution is verified; each host still needs the required runtime and supporting files.

    Read findings, coverage_complete and diff_performed together. Unsupported or unreadable inputs are reported explicitly. Exit zero means the report was produced, including reports with findings or incomplete coverage.

    New in Testing & Debugging