From Ratified Checklist to Working Tool: Redbelly DAO's Proposal Pre-Screening Framework

hildecorp — Mon, 17 Aug 2026 07:31:40 +0000

Redbelly Community DAO ratified an 11-criterion Proposal Review Checklist on 6 October 2025. It passed unanimously. It has been applied by hand ever since, which means it has been applied inconsistently. This is a writeup of a working pre-screening tool built on top of that ratified checklist, plus what got found along the way about gaps and changes in the DAO's own governing text.

The problem

The checklist itself was never in question. What was missing was enforcement. A proposer writes up a request, a Guild reviews it, the High Council reviews it again, and at no point does anything check the request against all 11 ratified criteria the same way twice. This project does not propose new criteria. It takes the 11 that already passed a vote and builds the missing enforcement layer.

What got built

Four pieces, in this order.

A rubric document maps all 11 criteria to a pass condition, a flag condition, and the specific Constitution section or Code of Conduct clause each one enforces, where one exists. Five of the 11 have no clean constitutional anchor. Four never had one. A fifth, Strategic Fit, lost its anchor partway through this build, when the Constitution itself changed under it. The rubric says so directly rather than inventing one.

A structured submission form collects a proposal in fields, not free text, one group of fields per criterion, so pre-screening has something consistent to check against.

A pre-screening flow runs a submitted proposal against all 11 rules and returns a verdict per criterion: pass or flag, the reason in plain language, and the citation. Every flag reason is a substance check against the ratified text: a missing field, an amount over a ratified cap, a group outside the ratified list. Never a length proxy for effort. Nothing here is a final decision. The output says so on every result: pre-screening runs before Guild and High Council review, not instead of it.

A demo mode runs the tool against two real, ratified proposals rather than invented examples.

What the ratified text does not agree with itself on

Building the rubric meant reading the checklist and the Constitution side by side, and the picture moved twice while this was in progress.

The checklist's Strategic Fit criterion names three working groups: Community, Marketing, Developers/Builders. Constitution v1.2 ratified five pods for this: Marketing, Builder/Develop, Researcher, Community, Partnerships. A DAO vote to consolidate the guild structure, passed 20 September 2025, merged 2 of those 5 pods into the other 3, leaving the same three the checklist had named all along. Then Constitution v1.3 removed the pod structure section from the Constitution entirely. The three groups the checklist names are still the operative ones in the DAO today. There is simply no Constitution text left to cite for them. Strategic Fit still checks a submission's pod against those three names, exactly as the ratified checklist states it. It just carries no citation link anymore. A flag here is the expected outcome for any submission, not a defect.

The eleventh listed criterion, Community Involvement, has no flag condition anywhere in the ratified text. Every other criterion pairs a pass condition with an explicit rule for when to flag it. This one does not. The tool treats a missing community-involvement field as informational only, never as a reason to mark a proposal as failing pre-screening, because inventing a flag condition the DAO never ratified would mean the tool enforcing a rule nobody voted on.

Both are stated plainly on the framework's own page rather than resolved quietly. A pre-screening tool that hides its own edge cases is worse than no tool.

Testing it against two real proposals

The demo runs two real FINPR marketing proposals, chosen because one passed and one failed.

The first, a single editorial press release with no influencer or promotional component, passed in March 2026 with 85.42 percent in favor. The second, a three-month renewal that added influencer marketing and social media management on top of the original scope, was rejected in July 2026 with 71.36 percent against.

Run the rejected proposal through the 11 criteria one by one and nothing obviously fails. The budget is split and justified. Payment is milestone-based. A reviewer is named. Risks are listed with mitigations. On the ratified checklist alone, this proposal does not trip a flag.

What actually sank it was not a checklist violation. It renewed and expanded a relationship on the strength of a narrower approval that never covered influencers or social media management in the first place. That is a precedent question, not a budget or milestone question, and the 11 ratified criteria do not test for it. Worth stating directly: an automated pre-screen built only on these 11 criteria would have let this proposal through clean, and the DAO still voted it down. The tool is honest about that limit rather than papering over it.

Sources

Every rule in the rubric traces back to one of these ratified documents. No summary, no paraphrase substituted for the original.

Ratified 11-criterion checklist, Snapshot proposal: https://snapshot.box/#/s:rbnt.eth/proposal/0xf2a05384e37a710c1600db1abbac9b4dc66444a56a1ed49df7f0e3dbfd7570e7

Constitution v1.3, current, supersedes v1.2 (ratified 20 Sep 2025): https://cdn.jsdelivr.net/gh/hildecorp/redbellydaotask25@main/website/constitutionv1.3.pdf

DAO Guild Structure Consolidation, passed 20 Sep 2025: https://snapshot.box/#/s:rbnt.eth/proposal/0x901e2873e7907d2ee87de5a79beae4fd602e75d053cd9fc6489e97d2547dd039

Code of Conduct v1.0, ratified: https://firebasestorage.googleapis.com/v0/b/redbelly-community-dao.firebasestorage.app/o/resources%2F1762476540292-Redbelly%20Community%20DAO%20Code%20of%20Conduct%20-%20Ratified.pdf?alt=media&token=84ca2c33-a502-4991-9720-4e6154847416

Full rubric, PDF: https://cdn.jsdelivr.net/gh/hildecorp/redbellydaotask25@main/website/TASK-25Proposal-Evaluation-Rubric.pdf

Full rubric, source repository: https://github.com/hildecorp/redbellydaotask25

Rejected worked example, proposal record: https://snapshot.box/#/s:rbnt.eth/proposal/0x21afed409a2b189d564dc83f9aac23e3f1c3f7c585d0c35e8a0826d3aa505d04

The passed worked example, Marketing Press Only with FINPR Agency, is referenced against the DAO's own governance changelog rather than linked directly here, since its individual proposal record was not captured firsthand during this build. Pull it directly from the Snapshot archive above before treating its 85.42 percent figure as final.

Live framework: https://redbellydaotask25.vercel.app/

DEV Community: hildecorp