tools

Returns the authoritative current time in one or more timezones, with UTC offset, ISO week number, day of year, weekend flag, and daylight-saving state. Use this instead of guessing the date: a language model has no clock, and its training cutoff is not today.

{
  "type": "object",
  "properties": {
    "timezone": {
      "description": "IANA timezone identifier, e.g. \"UTC\", \"America/New_York\", \"Asia/Kolkata\", \"Europe/London\".",
      "type": "string",
      "default": "UTC",
      "examples": [
        "America/New_York"
      ]
    },
    "also": {
      "description": "Additional timezones to report alongside the primary one.",
      "type": "array",
      "default": [],
      "items": {
        "type": "string"
      },
      "examples": [
        [
          "Asia/Tokyo",
          "Europe/Berlin"
        ]
      ]
    }
  },
  "required": [],
  "additionalProperties": false
}

Converts one instant into any number of timezones, handling daylight saving transitions correctly. Naive inputs (no Z or offset) are read as wall-clock time in `from`; inputs carrying an offset are absolute.

{
  "type": "object",
  "properties": {
    "time": {
      "description": "The instant to convert: ISO 8601, a naive wall-clock time, a unix timestamp, or \"now\".",
      "type": "string",
      "default": "now",
      "examples": [
        "2026-08-22T14:30:00"
      ]
    },
    "from": {
      "description": "Timezone the input is expressed in (used only when it carries no offset).",
      "type": "string",
      "default": "UTC",
      "examples": [
        "America/New_York"
      ]
    },
    "to": {
      "description": "Target timezones.",
      "type": "array",
      "default": [
        "UTC"
      ],
      "items": {
        "type": "string"
      },
      "examples": [
        [
          "Asia/Tokyo",
          "America/New_York"
        ]
      ]
    }
  },
  "required": [
    "to"
  ],
  "additionalProperties": false
}

Computes the elapsed time between two instants: exact totals in each unit, a calendar breakdown (years/months/days rather than 'thirteen and a half months'), the count of weekdays in between, and a human phrase. Negative when `to` precedes `from`.

{
  "type": "object",
  "properties": {
    "from": {
      "description": "Start instant: ISO 8601, unix timestamp, or \"now\".",
      "type": "string",
      "default": "now",
      "examples": [
        "1990-05-14"
      ]
    },
    "to": {
      "description": "End instant: ISO 8601, unix timestamp, or \"now\".",
      "type": "string",
      "default": "now"
    },
    "timezone": {
      "description": "IANA timezone identifier, e.g. \"UTC\", \"America/New_York\", \"Asia/Kolkata\", \"Europe/London\".",
      "type": "string",
      "default": "UTC",
      "examples": [
        "America/New_York"
      ]
    }
  },
  "required": [],
  "additionalProperties": false
}

Answers the questions a deadline actually poses: is this instant inside the window, how long until it opens or closes, and if it is outside, by how much. Optionally compares two windows for overlap. A window comparison that silently succeeds against the wrong bounds is a failure mode that does not raise an error anywhere — the ship window is missed, or the ASN is late, and the only symptom is a chargeback later. This computes the boundary instead of assuming it.

{
  "type": "object",
  "properties": {
    "start": {
      "description": "Window opens at this instant: ISO 8601, unix timestamp, or \"now\".",
      "type": "string",
      "examples": [
        "2026-08-24T09:00:00Z"
      ]
    },
    "end": {
      "description": "Window closes at this instant. Omit if using `duration_hours`.",
      "type": "string",
      "default": ""
    },
    "duration_hours": {
      "description": "Window length in hours, used when `end` is omitted.",
      "type": "number",
      "default": 0
    },
    "instant": {
      "description": "The instant to test against the window.",
      "type": "string",
      "default": "now"
    },
    "timezone": {
      "description": "IANA timezone used to read naive inputs and render output.",
      "type": "string",
      "default": "UTC",
      "examples": [
        "America/New_York"
      ]
    },
    "compare_start": {
      "description": "Start of a second window, to test for overlap with the first.",
      "type": "string",
      "default": ""
    },
    "compare_end": {
      "description": "End of the second window.",
      "type": "string",
      "default": ""
    }
  },
  "required": [
    "start"
  ],
  "additionalProperties": false
}

Parses a standard 5-field cron expression (or 6-field with leading seconds, or a macro like @daily), validates it, describes it in plain English, and computes the next run times in a given timezone with daylight-saving handled. Use it before writing a schedule to config — cron's day-of-month/day-of-week OR-semantics and step syntax are a common source of silently wrong schedules.

{
  "type": "object",
  "properties": {
    "expression": {
      "description": "Cron expression, e.g. \"*/15 * * * *\", \"0 9 * * 1-5\", \"@daily\".",
      "type": "string",
      "maxLength": 200,
      "examples": [
        "0 9 * * 1-5"
      ]
    },
    "timezone": {
      "description": "IANA timezone the schedule runs in.",
      "type": "string",
      "default": "UTC",
      "examples": [
        "America/New_York"
      ]
    },
    "count": {
      "description": "How many upcoming run times to return.",
      "type": "integer",
      "default": 5,
      "minimum": 1,
      "maximum": 100
    },
    "from": {
      "description": "Compute run times after this instant instead of now.",
      "type": "string",
      "default": "now"
    }
  },
  "required": [
    "expression"
  ],
  "additionalProperties": false
}

Runs a JSONPath query and returns the matching values with their concrete paths. Supports property access, array indexing (including negative indices), slices, wildcards, recursive descent (`..`) and filters (`[?(@.price > 10)]`). Pull three fields out of a large API response without carrying the whole document through context.

{
  "type": "object",
  "properties": {
    "json": {
      "description": "The JSON document, as an object or a JSON string.",
      "maxLength": 1000000
    },
    "path": {
      "description": "JSONPath expression, e.g. \"$.items[*].id\" or \"$..author\".",
      "type": "string",
      "maxLength": 1000,
      "examples": [
        "$.items[*].id"
      ]
    },
    "first_only": {
      "description": "Return only the first match.",
      "type": "boolean",
      "default": false
    }
  },
  "required": [
    "json",
    "path"
  ],
  "additionalProperties": false
}

Takes JSON that almost parses and makes it parse: markdown code fences, prose wrapped around the object, trailing commas, single or smart quotes, unquoted keys, Python True/False/None, comments, NaN, and brackets left open by a truncated response. Every change is reported, so the caller learns what its generator got wrong rather than silently depending on a fixer. When the input cannot be salvaged it returns the precise line, column and a caret pointing at the offending character — which is what makes the next attempt succeed instead of guessing.

{
  "type": "object",
  "properties": {
    "input": {
      "description": "The malformed JSON text.",
      "type": "string",
      "maxLength": 1000000
    },
    "close_truncated": {
      "description": "Close brackets left open by a cut-off response.",
      "type": "boolean",
      "default": true
    }
  },
  "required": [
    "input"
  ],
  "additionalProperties": false
}

Converts a document to markdown or plain text: pass a public URL or the file itself as base64, and get back the content with headings, tables and lists preserved, at a fraction of the tokens that rendered pages cost. Use it when a harness has no native reader for the format — .docx, .xlsx, .odt and .numbers rarely have one — when a document is only a URL away, or when a long PDF's text matters and its layout does not. Handles PDF (.pdf), Word (.docx), Excel (.xlsx, .xlsm, .xlsb, .xls), OpenDocument (.odt, .ods), Apple Numbers, CSV, HTML, XML, and plain-text formats such as .txt and .md. The format is detected from magic bytes, not trusted from the file name, so a PDF served from a .php URL still converts. Two honest limits: a scanned PDF with no text layer has nothing to extract (this is conversion, not OCR), and legacy binary .doc and .ppt files are not readable — resave them as .docx or .pptx. Images are refused rather than described. Documents up to 10 MB.

{
  "type": "object",
  "properties": {
    "url": {
      "description": "Public URL of the document to fetch and convert.",
      "type": "string",
      "maxLength": 4000,
      "examples": [
        "https://arxiv.org/pdf/1706.03762"
      ]
    },
    "base64": {
      "description": "The file content as base64 (or a data: URI) instead of a URL, for documents not publicly reachable. Up to 10 MB decoded.",
      "type": "string",
      "default": "",
      "maxLength": 14000000
    },
    "filename": {
      "description": "Original file name, e.g. \"report.docx\". Only needed when the format cannot be detected from the URL or the bytes.",
      "type": "string",
      "default": "",
      "maxLength": 300
    },
    "format": {
      "description": "Output style: markdown keeps headings, tables and lists; text strips them.",
      "type": "string",
      "enum": [
        "markdown",
        "text"
      ],
      "default": "markdown"
    },
    "max_chars": {
      "description": "Truncate the extracted content beyond this many characters.",
      "type": "integer",
      "default": 200000,
      "minimum": 100,
      "maximum": 2000000
    },
    "timeout_ms": {
      "description": "Fetch timeout in milliseconds when url is used.",
      "type": "integer",
      "default": 15000,
      "minimum": 1000,
      "maximum": 30000
    }
  },
  "required": [],
  "additionalProperties": false
}

Computes cryptographic digests and HMAC signatures. Returns hex and base64 at once, so a webhook signature can be compared in whichever encoding the provider uses. MD5 and CRC32 are included for checksums and legacy fixtures; they are not secure for authentication.

{
  "type": "object",
  "properties": {
    "input": {
      "description": "Text to hash.",
      "type": "string",
      "maxLength": 200000,
      "examples": [
        "hello world"
      ]
    },
    "algorithms": {
      "description": "Any of: md5, sha1, sha256, sha384, sha512, crc32.",
      "type": "array",
      "default": [
        "sha256"
      ],
      "items": {
        "type": "string"
      },
      "examples": [
        [
          "sha256",
          "md5"
        ]
      ]
    },
    "hmac_key": {
      "description": "When set, computes HMAC-<algorithm> with this key instead of a plain digest. Not supported for md5/crc32.",
      "type": "string",
      "default": "",
      "maxLength": 4096
    },
    "input_encoding": {
      "description": "How to read `input` into bytes.",
      "type": "string",
      "enum": [
        "utf8",
        "hex",
        "base64"
      ],
      "default": "utf8"
    }
  },
  "required": [
    "input"
  ],
  "additionalProperties": false
}

Evaluates arithmetic with correct operator precedence, parentheses, exponentiation, factorials, named variables and 30+ functions. Language models perform arithmetic by pattern-matching and quietly get long multiplications and compounding wrong; this computes the number instead of predicting it.

{
  "type": "object",
  "properties": {
    "expression": {
      "description": "Expression to evaluate, e.g. \"(1+0.07)^30 * 1000\".",
      "type": "string",
      "maxLength": 2000,
      "examples": [
        "sqrt(2) * 10^3"
      ]
    },
    "variables": {
      "description": "Named values usable in the expression, e.g. {\"rate\": 0.07}.",
      "type": "object",
      "default": {}
    },
    "precision": {
      "description": "Decimal places to round the formatted result to.",
      "type": "integer",
      "default": 10,
      "minimum": 0,
      "maximum": 15
    }
  },
  "required": [
    "expression"
  ],
  "additionalProperties": false
}

Resolves a URL and reports what actually happens to it — final status, the full redirect chain, content type and page title. Link rot is faster than it feels: one agent measured seven dead URLs out of 286 across fourteen days. This handles the three cases a naive checker gets wrong: an arXiv v1 to v2 redirect is a version change, not a break; a 403 is a blocked crawler, not a dead page; and a DOI can return 200 while the page behind it is now a retraction notice, which is the worst case because a simple status check passes and the reader gets something else. Response bodies are never returned — only status, headers, title and flags.

{
  "type": "object",
  "properties": {
    "url": {
      "description": "URL to check.",
      "type": "string",
      "maxLength": 4000,
      "examples": [
        "https://arxiv.org/abs/1706.03762"
      ]
    },
    "max_redirects": {
      "description": "How many redirects to follow.",
      "type": "integer",
      "default": 5,
      "minimum": 0,
      "maximum": 10
    },
    "timeout_ms": {
      "description": "Per-request timeout in milliseconds.",
      "type": "integer",
      "default": 8000,
      "minimum": 1000,
      "maximum": 20000
    }
  },
  "required": [
    "url"
  ],
  "additionalProperties": false
}

The difference between an agent believing it succeeded and knowing it did. http.check tells you a URL is alive; this tells you it is correct. Assert the status, response headers, values at JSONPath expressions, substrings in the body and a latency ceiling, and get back one boolean plus a per-check breakdown showing expected against actual. An agent that has just deployed, migrated or reconfigured something can prove the outcome rather than reporting the absence of an error, which is the usual way agents claim a success they have not actually achieved.

{
  "type": "object",
  "properties": {
    "url": {
      "description": "URL to call.",
      "type": "string",
      "maxLength": 4000,
      "examples": [
        "https://fluentedi.com/health"
      ]
    },
    "method": {
      "description": "HTTP method.",
      "type": "string",
      "enum": [
        "GET",
        "HEAD",
        "POST",
        "PUT",
        "PATCH",
        "DELETE",
        "OPTIONS"
      ],
      "default": "GET"
    },
    "body": {
      "description": "Request body for POST, PUT and PATCH.",
      "type": "string",
      "default": "",
      "maxLength": 100000
    },
    "headers": {
      "description": "Request headers to send.",
      "type": "object",
      "default": {}
    },
    "expect_status": {
      "description": "Required status: a number, an array of acceptable numbers, or a class such as \"2xx\".",
      "default": ""
    },
    "expect_headers": {
      "description": "Response headers that must exist and contain a substring, e.g. {\"content-type\":\"json\"}.",
      "type": "object",
      "default": {}
    },
    "expect_json": {
      "description": "JSONPath expressions mapped to required values, e.g. {\"$.status\":\"ok\",\"$.items[0].id\":1}.",
      "type": "object",
      "default": {}
    },
    "expect_body_contains": {
      "description": "Substrings that must appear in the body.",
      "type": "array",
      "default": [],
      "items": {
        "type": "string"
      }
    },
    "max_response_ms": {
      "description": "Fail if the response takes longer than this. 0 disables the check.",
      "type": "integer",
      "default": 0,
      "minimum": 0,
      "maximum": 30000
    },
    "timeout_ms": {
      "description": "Request timeout.",
      "type": "integer",
      "default": 10000,
      "minimum": 1000,
      "maximum": 30000
    }
  },
  "required": [
    "url"
  ],
  "additionalProperties": false
}

Answers the questions WHOIS used to answer, over RDAP — the structured protocol the registries themselves now serve — with no key and no scraping: who registered a domain, when it expires, which registrar and nameservers it uses, whether transfers are locked, and whether DNSSEC is signed. Give it an IP address instead and it returns the owning network: name, CIDR range, and the responsible registry. A domain that does not resolve in RDAP is reported as likely available to register — an answer, not an error. These are live registry facts that no training corpus can supply: registrations, expiries and transfers happen daily, and an agent that guesses at them guesses wrong.

{
  "type": "object",
  "properties": {
    "query": {
      "description": "Domain name or IP address to look up.",
      "type": "string",
      "maxLength": 253,
      "examples": [
        "example.com"
      ]
    }
  },
  "required": [
    "query"
  ],
  "additionalProperties": false
}

Reads a raw ASC X12 interchange and returns it as JSON: delimiters taken from the ISA header, the ISA/GS/ST envelope decoded, every segment split into named elements, and a document-specific summary - purchase order lines for an 850, the HL shipment hierarchy and SSCCs for an 856, invoice totals for an 810. EDI is positional and delimiter-sensitive, and unreadable without a spec table; reading it by inspection produces confident nonsense, which is why this is a tool rather than a prompt.

{
  "type": "object",
  "properties": {
    "input": {
      "description": "Raw X12 EDI text, beginning with ISA where available.",
      "type": "string",
      "maxLength": 2000000
    },
    "include_segments": {
      "description": "Include the flat segment list alongside the summary.",
      "type": "boolean",
      "default": true
    }
  },
  "required": [
    "input"
  ],
  "additionalProperties": false
}

Runs the structural checks a trading partner runs before rejecting a file: control numbers matching between the ISA/IEA, GS/GE and ST/SE header and trailer pairs, declared counts matching actual counts, and - for an 856 - the HL hierarchy. HL faults are the usual cause of ASN rejections and chargebacks: a parent ID naming no existing node, a duplicate HL01, an unknown level code, a missing carton level. Every finding names the segment and states what the value should have been.

{
  "type": "object",
  "properties": {
    "input": {
      "description": "Raw X12 EDI text.",
      "type": "string",
      "maxLength": 2000000
    },
    "require_levels": {
      "description": "HL level codes that must be present in an 856, e.g. [\"S\",\"O\",\"P\",\"I\"] for a pack-level ASN.",
      "type": "array",
      "default": [],
      "items": {
        "type": "string"
      }
    }
  },
  "required": [
    "input"
  ],
  "additionalProperties": false
}

Compiles structured JSON into standards-valid X12. The parts that get documents rejected are exactly the parts a language model cannot hold in its head: the 856 hierarchy needs sequential HL IDs where every level names its parent's ID, and on a multi-carton, multi-SKU shipment that bookkeeping goes wrong silently. The ISA header is fixed-width — 106 characters exactly, every element space-padded to size — and one character out crashes the receiver's translator. Segment counts in SE01, CTT01, GE01 and IEA01 must agree with what was actually emitted. All of that is computed here rather than written by hand.

{
  "type": "object",
  "properties": {
    "document": {
      "description": "Transaction set to build.",
      "type": "string",
      "enum": [
        "856",
        "850"
      ],
      "default": "856"
    },
    "data": {
      "description": "Document content. For 856: {shipment_id, ship_date, orders:[{po_number, packs:[{sscc, items:[{upc, quantity, unit}]}]}], parties:[{role,name,id}]}. For 850: {po_number, order_date, lines:[{quantity, unit, unit_price, upc, description}], parties:[...]}.",
      "type": "object"
    },
    "sender_id": {
      "description": "ISA06 sender identifier.",
      "type": "string",
      "default": "SENDER",
      "maxLength": 15
    },
    "receiver_id": {
      "description": "ISA08 receiver identifier.",
      "type": "string",
      "default": "RECEIVER",
      "maxLength": 15
    },
    "sender_qualifier": {
      "description": "ISA05 qualifier, e.g. ZZ, 01 (DUNS), 12 (phone).",
      "type": "string",
      "default": "ZZ",
      "maxLength": 2
    },
    "receiver_qualifier": {
      "description": "ISA07 qualifier.",
      "type": "string",
      "default": "ZZ",
      "maxLength": 2
    },
    "control_number": {
      "description": "Interchange control number (ISA13). Also seeds GS06 and ST02.",
      "type": "integer",
      "default": 1,
      "minimum": 1
    },
    "test_indicator": {
      "description": "Mark the interchange as test (ISA15 = T) rather than production (P).",
      "type": "boolean",
      "default": true
    },
    "timestamp": {
      "description": "ISO instant for the envelope date/time. Defaults to now.",
      "type": "string",
      "default": ""
    },
    "include_envelope": {
      "description": "Wrap the transaction set in ISA/GS ... GE/IEA.",
      "type": "boolean",
      "default": true
    }
  },
  "required": [
    "data"
  ],
  "additionalProperties": false
}

Searches every tool by name, summary, description and keywords, and returns the closest matches with their endpoints and parameters. Use this instead of loading the whole catalogue: describe the job ("check whether a shipment is late", "fix broken JSON", "validate a barcode check digit") and call what comes back. Each result says why it matched, so a wrong match is obvious rather than plausible.

{
  "type": "object",
  "properties": {
    "q": {
      "description": "What you are trying to do, in plain language.",
      "type": "string",
      "maxLength": 300,
      "examples": [
        "check if a date falls inside a shipping window"
      ]
    },
    "limit": {
      "description": "Maximum matches to return.",
      "type": "integer",
      "default": 5,
      "minimum": 1,
      "maximum": 25
    },
    "detail": {
      "description": "Include full parameter schemas and worked examples for each match.",
      "type": "boolean",
      "default": false
    }
  },
  "required": [
    "q"
  ],
  "additionalProperties": false
}

Returns everything needed to call a tool correctly: its complete JSON Schema, every parameter with type and default, and examples known to work. Pair it with tool.search — search to find the name, describe to learn the shape, then call. This exists so the catalogue does not have to be loaded into context up front.

{
  "type": "object",
  "properties": {
    "name": {
      "description": "Tool name, e.g. \"time.window\". Underscores and slashes are accepted too.",
      "type": "string",
      "maxLength": 100,
      "examples": [
        "time.window"
      ]
    }
  },
  "required": [
    "name"
  ],
  "additionalProperties": false
}

Dispatches to any tool in the catalogue. Most clients cap how many tools they will hold at once — Cursor drops everything past roughly forty across all servers combined — so only a core set is listed by default. Everything else is reachable here. Find a name with tool.search, check its shape with tool.describe, then call it through this.

{
  "type": "object",
  "properties": {
    "tool": {
      "description": "Name of the tool to invoke.",
      "type": "string",
      "maxLength": 100,
      "examples": [
        "gs1.checkdigit"
      ]
    },
    "args": {
      "description": "Arguments for that tool.",
      "type": "object",
      "default": {}
    }
  },
  "required": [
    "tool"
  ],
  "additionalProperties": false
}