- Home
- Skills
- APIs & Backend
- API Breaking-Change Detector — CI Guardrail
Works with the AI Agents you already use
API Breaking-Change Detector — CI Guardrail
The skill also walks a seven-category ERR checklist on every run, proving the negative when no breaking changes are found, and explicitly states the CI exit code.
$19
API Breaking-Change Detector — CI Guardrail
Example session with this skill installed
Compare openapi-v1.yaml and openapi-v2.yaml. Flag breaking changes and generate all artifacts. The new version removes the 'email' field from the User schema response.
- Read your context and instructions
- Compiled the api breaking-change detector
Analysis complete. A breaking change was detected in the User schema. All artifacts have been generated.
API Contract Diff — v1.0.0 → v1.1.0
Engine: manual-read
Base: openapi-v1.yaml (v1.0.0)
New: openapi-v2.yaml (v1.1.0)
Summary
| Severity | Count | Notes |
|---|---|---|
ERR | 1 | Removed response field in User schema |
WARN | 0 | none |
INFO | 0 | none |
Exit code: 1 — breaking changes detected.
Breaking changes (ERR)
GET /users/{id}— Removed response fieldemailfrom theUserschema. This will break clients expecting this property.- Remediation: Restore the
emailfield or mark it as deprecated if a new field replaces it.
- Remediation: Restore the
ERR checklist
| Category | Result |
|---|---|
| Deleted endpoint | pass |
| Modified path parameter | pass |
| Status code shift | pass |
| Removed response field | fail |
| New required request field | pass |
| Type narrowing | pass |
| Enum value removal | pass |
API_CHANGELOG.md
# API Changelog
## [1.1.0]
### Breaking Changes
- Removed `email` property from `User` response object in `GET /users/{id}`.
diff.json
{
"schema_version": "1.0",
"engine": "manual-read",
"summary": { "err": 1, "warn": 0, "info": 0 },
"exit_code": 1,
"breaking": [
{
"id": "ERR-001",
"rule": "removed-response-field",
"location": "GET /users/{id}",
"field": "User.email",
"description": "Removed property 'email' from response.",
"remediation": "Restore field or start deprecation cycle."
}
],
"warnings": [],
"info": [],
"unchanged": ["POST /users", "GET /health"],
"err_checklist": {
"removed_response_field": "fail"
}
}
Next steps
- Review
BREAKING_CHANGES.mdfor full remediation details. - Revert the field removal or update the
ignore_fileif this is an intentional break. - Run the diff again to verify the fix.
Connects securely to your tools. The creator never sees your data.
About this skill
API Breaking-Change Detector is a Claude Code skill for detecting backward-incompatible changes between OpenAPI 3.x specifications. It performs semantic contract diffing on JSON or YAML specs and classifies every difference into three severity levels: ERR (breaking), WARN (non-breaking risk), and INFO (additive or metadata). The skill catches deleted endpoints, modified path parameters, removed response fields, new required request fields, type narrowing, and enum value removal—the exact changes that break existing API clients.
Every run produces a complete artifact set in a single pass: API_CHANGELOG.md, BREAKING_CHANGES.md, diff.json (schema-conformant, with breaking, warnings, and info arrays always present), and—when a deprecation or approved break exists—DEPRECATION_PLAN.md. The skill also walks a seven-category ERR checklist on every run, proving the negative when no breaking changes are found, and explicitly states the CI exit code.
Use cases: pull request gating, contract drift audits, version comparison before release, and automated changelog generation. Compatible with oasdiff, api-governor, and OpenAPI Comparator, or fully functional as a manual-read fallback when no tool is available.
Keywords: OpenAPI breaking change detector, API contract diff, backward compatibility checker, CI/CD API guardrail, semantic diff OpenAPI, contract drift detection, oasdiff skill.
Requires valid OpenAPI 3.x specs. Cannot resolve external $ref pointers if files are missing. Scope limited to contract structure, not runtime behavior.
How to use
- 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
Sign in to Agensi
Your agent shows a sign-in link. Sign in and click Allow.
- 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 1 month ago