jeff-adapter-is-complete

Document completeness check. Checks a set of documents against a checklist. Returns complete, incomplete or unclear, accounting for missing, expired, unsigned or incorrectly attributed documents.

A LoRA adapter for jeff-base v1.4, Jeff's small open decision model (a fine-tune of Qwen3.5-0.8B). You give it a situation and a question with named options; it returns a calibrated probability for every option from one forward pass, with no generated text to parse.

• Adapter ID: jeff.reasoning.is-complete, version 1.4.0; short names: oss-is-complete, is-complete
• Catalogue page: jeffhub.ai/catalogue/jeff.reasoning.is-complete
• Code, server and SDKs: github.com/firelex/jeff
• Base it needs: mstrasser/jeff-base at revision v1.4. The engine refuses an adapter trained on another base, so this adapter loads only on that revision.

When to use it

• You need to assess whether a document submission meets a checklist.

When not to use it

• You need an itemised list of missing documents; this adapter assesses the submission as a whole.

What it decides

What a call sends:

Field	Required	What it holds
context (context)	yes	The review: an object with the organisation, the reviewer, the kind of pack, its purpose and the date of the review.
checklist (context)	yes	The checklist: a list of required documents and conditions (some only for some cases).
submission (input, fields)	yes	The pack received: an object with the case (name, reference) and the documents received, each described (kind, date, pages, signatures).

jeff.reasoning.is-complete.is_complete (classify, choice)

Instructions: Is the pack complete against the checklist? Go through every checklist line that applies to this case. A required document counts only if it is the kind of document listed (not a similar one), is about the subject of the case, has all its pages, is signed where the checklist asks for a signature, has not expired on the date of the review, and was issued within the period the checklist allows. A conditional line applies only when its circumstance holds for the case. complete: every applicable line is met (extra documents do not matter). incomplete: at least one applicable line is not met. unclear: no line is clearly unmet, but at least one cannot be decided from what is shown, for example a needed date, signature or page count is not visible, a document's kind cannot be told, or the case details do not say whether a conditional line's circumstance holds while its document is absent. Cover notes and labels may be misleading; judge the documents themselves.

Labels:

• complete: Every document the checklist requires for this case is in the pack and acceptable.
• incomplete: At least one required document is missing or not acceptable (wrong kind, about someone or something else, expired or too old, missing pages, or unsigned).
• unclear: The pack and the case details do not settle it: something the answer depends on cannot be seen or is not stated.

How it was measured

Base: Jeff v1.4; adapter trained 2026-10-25.

Test	Rows	Accuracy
the in-house test (generator)	1,974	96.8%
a second in-house test of unseen cases (generator)	1,932	95.7%

Accuracy is the share of test rows whose top answer was right. Kinds: public, a public benchmark; real, held-out real documents; generator, an in-house test from the same generator as the training data (it measures less than a public or real test does); mixed, several kinds; unknown, not recorded.

When it is confident (from the calibration set): the threshold is the confidence from which an answer counts as confident, precision the share of confident answers that were right, coverage the share of answers that were confident.

Function	Threshold	Precision	Coverage	Calibration rows
is_complete	0.789	98.0%	96.7%	1,941

How to call it

Install the Python SDK and the engine (see the install guide; until the packages are on PyPI, install them from github.com/firelex/jeff).

The example below is "Bakery's business loan pack". The bakery has two directors but the pack holds photo identification for only one, so it is incomplete.

On your machine (Jeff.local downloads this adapter and the base from Hugging Face the first time):

import jeff

client = jeff.Jeff.local(["is-complete"])

decide = client.function("jeff.reasoning.is-complete.is_complete", context={
    "context": {
        "organisation": "A small business lender",
        "reviewer": "Loan operations analyst",
        "pack_kind": "Business loan application",
        "purpose": "Check the application pack is complete before credit review",
        "review_date": "2 November 2026",
    },
    "checklist": [
        "Signed loan application form (all pages)",
        "Business bank statements for the last three months",
        "Photo identification of every director",
        "Latest annual accounts (only for businesses trading more than two years)",
    ],
})

decision = decide({
    "case": {
        "name": "Ferngrove Bakery Ltd",
        "reference": "APP-20931",
        "trading_since": "March 2025",
        "directors": [
            "Priya Natarajan",
            "Tom Okafor",
        ],
    },
    "documents_received": [
        {
            "kind": "Loan application form",
            "date": "28 October 2026",
            "pages": "4 of 4",
            "signatures": "signed by Priya Natarajan",
        },
        {
            "kind": "Business bank statements",
            "date": "July, August and September 2026",
            "pages": "complete",
        },
        {
            "kind": "Passport",
            "date": "valid until 2031",
            "about": "Priya Natarajan",
        },
    ],
})

print(decision.value, decision.confidence)

Through the Jeff API (an API key from jeffhub.ai):

import os

import jeff

with jeff.Jeff.remote("https://api.jeffhub.ai", token=os.environ["JEFF_API_KEY"], timeout_s=30) as client:
    decide = client.function("jeff.reasoning.is-complete.is_complete", context={
        "context": {
            "organisation": "A small business lender",
            "reviewer": "Loan operations analyst",
            "pack_kind": "Business loan application",
            "purpose": "Check the application pack is complete before credit review",
            "review_date": "2 November 2026",
        },
        "checklist": [
            "Signed loan application form (all pages)",
            "Business bank statements for the last three months",
            "Photo identification of every director",
            "Latest annual accounts (only for businesses trading more than two years)",
        ],
    })

    decision = decide({
        "case": {
            "name": "Ferngrove Bakery Ltd",
            "reference": "APP-20931",
            "trading_since": "March 2025",
            "directors": [
                "Priya Natarajan",
                "Tom Okafor",
            ],
        },
        "documents_received": [
            {
                "kind": "Loan application form",
                "date": "28 October 2026",
                "pages": "4 of 4",
                "signatures": "signed by Priya Natarajan",
            },
            {
                "kind": "Business bank statements",
                "date": "July, August and September 2026",
                "pages": "complete",
            },
            {
                "kind": "Passport",
                "date": "valid until 2031",
                "about": "Priya Natarajan",
            },
        ],
    })

    print(decision.value, decision.confidence)

curl -s https://api.jeffhub.ai/v1/decide \
  -H "Authorization: Bearer $JEFF_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{"calls": [{"function": "jeff.reasoning.is-complete.is_complete", "input": {"case": {"name": "Ferngrove Bakery Ltd", "reference": "APP-20931", "trading_since": "March 2025", "directors": ["Priya Natarajan", "Tom Okafor"]}, "documents_received": [{"kind": "Loan application form", "date": "28 October 2026", "pages": "4 of 4", "signatures": "signed by Priya Natarajan"}, {"kind": "Business bank statements", "date": "July, August and September 2026", "pages": "complete"}, {"kind": "Passport", "date": "valid until 2031", "about": "Priya Natarajan"}]}, "context": {"context": {"organisation": "A small business lender", "reviewer": "Loan operations analyst", "pack_kind": "Business loan application", "purpose": "Check the application pack is complete before credit review", "review_date": "2 November 2026"}, "checklist": ["Signed loan application form (all pages)", "Business bank statements for the last three months", "Photo identification of every director", "Latest annual accounts (only for businesses trading more than two years)"]}}]}'

Recorded answer (2026-11-03): incomplete with confidence 0.965.

To download the weights yourself:

hf download mstrasser/jeff-base --revision v1.4 --local-dir jeff-base
hf download mstrasser/jeff-adapter-is-complete --revision v1.4 --local-dir adapters/is-complete

Files

File	What it is
adapter_model.safetensors, adapter_config.json	The LoRA weights and their configuration
readout.safetensors	The adapter's readout over the answer codes
decision_config.json	Answer codes, temperature, prompt layout and the base it was trained on (the SHA-256 of the base's weights, which the engine checks)
manifest.yaml	The adapter's manifest: its functions, labels, fields and calibration, as the SDKs read it
test.jsonl	The held-out test set the numbers above come from
calibration.jsonl	The calibration set the thresholds come from
LICENSE	The Apache License 2.0, for the Qwen3.5-0.8B notice below

Licence

Adapter licence: Apache-2.0

Qwen3.5-0.8B notice: these weights were modified from Qwen3.5-0.8B by the Jeff project: jeff-base is a fine-tune of Qwen3.5-0.8B, and this adapter was trained on top of it. Qwen3.5-0.8B is Copyright 2026 Alibaba Cloud and licensed under the Apache License, Version 2.0; a copy of that licence is in LICENSE.

Jeff is an independent project, not affiliated with or endorsed by the makers of Qwen.