JSON Repair

Overview

You send text that was meant to be JSON and we fix the mechanical things that keep it from parsing: trailing commas, single or curly quotes, unquoted keys, comments, Python's True and None, a markdown code fence, a JSONP callback wrapper, an ellipsis someone left in place of the rest of a list, and a document that simply stops mid-way. You get back canonical JSON plus a list naming each fix. We never change what your strings say and never invent a value, so if the text cannot be repaired by rule we tell you instead of guessing. Accepts file input via file_ref — no need to extract file content first.

Models and lenient tools produce almost-JSON all day long, and a strict parser rejects the whole document over one stray comma. Fixing it by rule is faster and far safer than asking a model to rewrite the payload, which risks changing your data.

When to use it

Scope

Non-goals

• We never check the result against a schema — that is schema.validate_json
• We never work out what the document should look like — that is schema.infer_json_schema
• We never repair broken character encoding or scrambled text
• We never guess a missing value; unrepairable input is reported, not invented

Properties

Capabilities

Cost & billing

Flat rate per invocation

File handling

File limits

Constraints this skill applies to your input and generated files. The “Limited by” column shows the static source of each cap; none of these can be raised by upgrading your plan.

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":"schema.repair_json","input":{"json":"{\"a\": 1,}"}}'

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

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

const { output } = await sq.execute("schema.repair_json", {"json":"{'name': 'Alice', 'roles': ['admin', 'user',],}"});

Errors

Other limits

Limitations

• We never check the result against a schema — that is schema.validate_json
• We never work out what the document should look like — that is schema.infer_json_schema
• We never repair broken character encoding or scrambled text
• We never guess a missing value; unrepairable input is reported, not invented
• Checking data against a schema (schema.validate_json)
• Working out a schema from examples (schema.infer_json_schema)
• Recovering text that is corrupted or in the wrong encoding

How to call

POST /v1/execute (requires authentication)

{
  "skillKey": "schema.repair_json",
  "input": {
    "json": "{'name': 'Alice', 'roles': ['admin', 'user',],}"
  }
}

Examples

{
  "json": "{'name': 'Alice', 'roles': ['admin', 'user',],}"
}

{
  "json": "{\"name\":\"Alice\",\"roles\":[\"admin\",\"user\"]}",
  "changed": true,
  "repairs": [
    "converted_single_quotes",
    "removed_trailing_comma"
  ]
}

{
  "json": "```json\n{\"a\": 1}\n```"
}

{
  "json": "{\"a\":1}",
  "changed": true,
  "repairs": [
    "removed_code_fence"
  ]
}

Recommended use cases

• Fix output from a lenient tool that left trailing commas or comments behind
• Clean up a model's answer: drop the surrounding code fence and close a reply that stopped mid-object or mid-string
• Drop the ellipsis a model used to abbreviate a long list
• Unwrap a callback envelope back into plain JSON
• Clear another skill's malformed-JSON failure before you retry it
• Bring a hand-edited config file back to strict JSON

Avoid for

• Checking data against a schema (schema.validate_json)
• Working out a schema from examples (schema.infer_json_schema)
• Recovering text that is corrupted or in the wrong encoding

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