Snapshot Diff

Overview

You send the snapshot you saved and the one you just produced, as text or as JSON, and we show you the difference the way a code review would: hunks with a few lines of context, each line marked as kept, added, or removed. First we tidy both sides the same way, using only the steps you ask for — drop the fields that change every run, settle on one line ending, trim trailing spaces — and JSON is put in a canonical shape so key order never shows up as a change. You also get a hash of each tidied side. Nothing is saved and no change is ever approved for you.

Snapshot tests fail for two reasons — something really changed, or a timestamp moved — and telling those apart by eye is exactly the work you wanted automated.

When to use it

Scope

Non-goals

• Does not write or update a stored snapshot
• Does not decide that a change is acceptable
• Does not judge whether two different wordings mean the same thing
• Does not run pattern replacements you invent — only the three tidying steps named here

Properties

Capabilities

Cost & billing

Flat rate per invocation

File handling

How we handle your data

Input schema

Output schema

API contract

curl -sS -X POST "$BASE/v1/execute" -H "Authorization: Bearer $KEY" -H 'Content-Type: application/json' -d '{"skillKey":"quality.snapshot_diff","input":{"expected":{"type":"text","value":"a"},"actual":{"type":"text","value":"a"}}}'

import { Squillify } from "@squillify/sdk";

const sq = new Squillify({ apiKey: process.env.SQUILLIFY_API_KEY });

const { output } = await sq.execute("quality.snapshot_diff", {"actual":{"type":"text","value":"a\nc"},"expected":{"type":"text","value":"a\nb"}});

Errors

Other limits

Limitations

• Does not write or update a stored snapshot
• Does not decide that a change is acceptable
• Does not judge whether two different wordings mean the same thing
• Does not run pattern replacements you invent — only the three tidying steps named here
• Saving or approving the new snapshot
• Comparing prose for meaning rather than characters

How to call

POST /v1/execute (requires authentication)

{
  "skillKey": "quality.snapshot_diff",
  "input": {
    "actual": {
      "type": "text",
      "value": "a\nc"
    },
    "expected": {
      "type": "text",
      "value": "a\nb"
    }
  }
}

Examples

{
  "actual": {
    "type": "text",
    "value": "a\nc"
  },
  "expected": {
    "type": "text",
    "value": "a\nb"
  }
}

{
  "equal": false,
  "appliedNormalizers": [],
  "_meta": {
    "granularity": "line",
    "diffProfile": "text-diff-1",
    "truncated": false
  },
  "summary": {
    "expectedHash": "7e18f737311b2dc3b2f269dd78396b0351f14fb66efa879f768cb23181883c78",
    "added": 1,
    "changed": 1,
    "removed": 1,
    "actualHash": "9e58d7137c654f526a7a7c9cbab79c2e859b4dfbb579d1d6dd3aa4113a8a909b"
  },
  "hunks": [
    {
      "newCount": 2,
      "oldCount": 2,
      "lines": [
        {
          "value": "a",
          "op": "context"
        },
        {
          "value": "b",
          "op": "remove"
        },
        {
          "value": "c",
          "op": "add"
        }
      ],
      "newStart": 1,
      "oldStart": 1
    }
  ]
}

{
  "actual": {
    "type": "text",
    "value": "ok"
  },
  "expected": {
    "type": "text",
    "value": "ok"
  }
}

{
  "equal": true,
  "appliedNormalizers": [],
  "_meta": {
    "granularity": "line",
    "diffProfile": "text-diff-1",
    "truncated": false
  },
  "summary": {
    "expectedHash": "2689367b205c16ce32ed4200942b8b8b1e262dfc70d9bc9fbc77c49699a4f1df",
    "added": 0,
    "changed": 0,
    "removed": 0,
    "actualHash": "2689367b205c16ce32ed4200942b8b8b1e262dfc70d9bc9fbc77c49699a4f1df"
  },
  "hunks": []
}

Recommended use cases

• Fail a snapshot test and show the reviewer exactly what moved
• Compare a golden file while ignoring the timestamp inside it
• Store the hash of a tidied snapshot and notice when output drifts

Avoid for

• Saving or approving the new snapshot
• Comparing prose for meaning rather than characters

Related skills

Composable with

Safety & privacy

• Runs without calling any external service.
• Your payloads are never stored permanently.
• Telemetry captures metadata only — never your payload contents.
• We never train models on your payloads.
• No human reviews your payloads.

Tags