LodgeiT .ORG

← Bounties  /  Bounty #15

Legal opinion · OPEN · Seeking reviewer $2,400 ex GST

Legal read on paying practitioners per use of a rule they have signed

A design-stage legal opinion. Two questions, each answerable yes or no with written reasons — on tax-agent reliance and the shape of a signer’s reward. Not a treatise, not advice on a client matter.

Legal opinion — new bounty type ~5 hrs Australian legal practitioners

Context: how LodgeiT writes tax law as Logical English, and Bounty #14, the rule-review bounty whose signer-payment model these questions test.

What we are asking

Two questions, each answerable yes or no with written reasons. We are not asking for a treatise, a redraft of our model, or advice on a specific client matter.

Q1. TASA s 90-5 reliance. If a registered tax agent's identity and credential are published in a public registry against a static, content-addressed rule file (identified by its SHA-256) and its test suite, while (a) the agent's identity is omitted from every individual calculation response, (b) each response carries only the rule hash, reasoner version, effective period and trust tier, and (c) the only contract is between the platform and the API caller, is the agent sufficiently insulated from "providing a tax agent service" to the end caller within the meaning of s 90-5 of the Tax Agent Services Act 2009 (Cth)?

Q2. Payment shape. Given that the payment to the agent is identical whether the computed result is a nil gain, a ten-million-dollar gain or a refusal (so it never tracks a taxpayer's outcome), does a per-invocation royalty accrued into a conditional vesting escrow (vesting at period end if no upheld defect finding against that hash; accepted finder bounties deducted before vesting; no clawback of amounts already paid) create any regulatory, Code of Professional Conduct, or professional-indemnity friction that a quarterly usage-banded licence fee for a software quality-assurance service, with a liquidated-damages warranty holdback, would avoid? If yes, name the friction and the provision or policy term it arises under.

Background (enough to answer; the full design note is available on request)

LodgeiT publishes tax rules written in Logical English, a controlled English that a logic reasoner runs directly. Each rule file is identified by its SHA-256. A rule file is deployed behind a paid API: a caller pays a small toll per query, supplies facts, and receives a figure with the chain of provisions that produced it, or a refusal naming the missing fact or the reason the scenario is out of scope.

Practitioners are paid to review rule files through public bounties (see Bounty #14 for the current example). A reviewer is paid a flat fee for the review regardless of verdict, and a graded bounty for each accepted defect. A reviewer who also signs the file as stating the law is proposed to receive, in addition, a share of the toll on every query answered by that exact hash for its effective period. The share is held in escrow and vests conditionally; defects found later are paid to the finder from it. The agent's name is bound to the hash in a public registry; the caller's response carries the hash and a trust tier (unreviewed / reviewed / ratified / disputed), not the name.

The design intent is that the signer's reward rewards being right, and that the finder's reward is funded by the signer who was wrong. The two questions above are the points where that design touches tax-agent regulation and the professional-conduct rules.

Deliverable

A written memorandum, in your own words, of 1,500 to 3,000 words, containing:

  1. A one-line answer to each question (yes / no / yes-with-conditions), stated first.
  2. Reasons, citing the provisions, Code items, APES 110 paragraphs, TPB guidance, insurance terms or authorities relied on. Where you rely on an authority, give enough to find it.
  3. For any "yes-with-conditions", the conditions as a numbered list we can implement or send to a drafter.
  4. Anything in the background that you consider a worse exposure than the two we asked about, in one paragraph. (We will pay a further $300 ex GST for each such point we accept as material.)
  5. An attestation (below).

Format: Markdown or PDF. Submitted via the bounty process; the memorandum is published on the bounty page with your name unless you ask for it to be published anonymously, in which case your credential is verified by LodgeiT and recorded as "verified Australian legal practitioner".

Attestation

The memorandum must include, in your words, statements that: you hold a current Australian practising certificate (state which); the reasoning is your own; if you used an AI tool at any stage, which tool and for what (drafting, research, checking), and that you verified every citation yourself. Use of AI is not disqualifying; undisclosed use is.

Acceptance

The bounty is paid when the memorandum answers both questions with reasons tied to identifiable provisions or authorities. It is paid whether the answers are favourable to the design or not; an answer of "no, and here is why the design cannot work as described" earns the full fee. Disagreement with our design is the point of asking.

What this is not

This is a design-stage opinion for LodgeiT's internal use in deciding how to structure the model. It is not advice to any taxpayer, practitioner or caller, and LodgeiT will not represent it as such. If you want the memorandum to carry a reliance limitation or a disclaimer of retainer, include it; that does not affect acceptance.

How to claim

Through the bounty page process: register interest, confirm your practising certificate, receive the full design note (v0.2) if you want it, submit within 14 days of acceptance.

Register interest

Claim Bounty #15

Register interest and confirm your practising certificate, then submit within 14 days of acceptance. Email support@lodgeit.net.au.

support@lodgeit.net.au