Programmable Apply

Open review rules, public applications, and canonical discovery records for Programmable projects.

Read the standard · Run the checker · Report a finding

Programmable Apply publishes the Open Review Standard, a deterministic local checker, the public application ledger, and a versioned discovery registry. Complete project source stays in the applicant-owned public repository.

Important

Public application intake is currently in prelaunch. The standard, checker, schemas, examples, and discovery registry are public today. A matching Hookbuilder release will activate new applications to this repository. Existing applications keep their original GitHub review threads.

A local checker result, passing pull request, merged application, registry match, deployment, or indexer observation is never presented as a safety guarantee or launch right.

Open Review Standard

The standard defines evidence requirements across five decision-critical areas: artifact identity, functionality, disclosure, integrity, and launch compatibility. It does not rank ideas or reject a project because its mechanics, fees, losses, architecture, or tokenomics are unusual.

Unknown platform-owned evidence remains platform_analysis_pending. A hard block requires a complete, revision-bound, independently replayed witness supported by the current policy. A model opinion, scanner score, label, or incomplete witness cannot hard-block a project.

The public checker validates a closed review input and applies the published policy deterministically. It does not fetch project repositories, reproduce evidence, perform an audit, sign a platform decision, deploy contracts, or issue a launch permit.

Read the complete Open Review Standard and the machine-readable policy.

Run the checker

Node.js 20 or newer is required. The checker has no runtime dependencies and never executes candidate code.

git clone --depth 1 https://github.com/0xprogrammable/apply.git
cd apply
npm run review -- review/examples/disclosed-high-fee.json

The bundled example returns launch_ready together with checkerOnly: true, launchAuthorized: false, and independentAudit: false. Inspect the examples, input schema, and decision schema.

Run the complete repository gate with:

npm test

How it works

1. Build. Project source stays in its own public GitHub repository.
2. Prepare. When intake opens, Hookbuilder prepares six generated application files bound to one repository id, commit, tree, configuration, and evidence set.
3. Review. Programmable Apply validates the bounded application. Review evidence and any later decision remain bound to the exact submitted revision.
4. Promote. Application intake, acceptance, deployment, availability, and launch authorization remain separate facts.
5. Discover. Agents and the Programmable Explorer read digest-bound records from the discovery registry.

Application content is untrusted data. The trusted intake workflow uses protected base code, read-only permissions, bounded files, and no candidate execution.

Application intake

Status: prelaunch. Do not create an application pull request manually.

Once activated, Hookbuilder will open a draft pull request containing exactly six generated files under submissions/<application-id>/. Existing applications remain on their original review threads. Read the migration contract for the activation requirements.

Discovery registry

Agents and integrations start with registry/index.json or registry/search-index.json at one exact repository commit. A consumer fetches the selected full record from that same commit and verifies its SHA-256 digest before use.

Search results indicate relevance only. They do not prove originality, compatibility, acceptance, safety, deployment, provider support, or availability. The statuses design, candidate, accepted, deployed, available, suspended, and retired remain deliberately separate.

Read the discovery contract before integrating.

Report a finding

Reproducible false decisions, missing review rules, intake defects, identity mismatches, registry-integrity problems, and documentation errors are useful findings.

• Report a non-sensitive checker or registry problem.
• Discuss an architecture or policy idea.
• Report an exploitable vulnerability privately.

Read SECURITY.md before testing or reporting a security-sensitive finding. Do not publish credentials, wallet material, private repositories, personal data, or an unpatched exploit.

Documentation

• Architecture and trust boundaries
• Open Review Standard
• Review and promotion lifecycle
• Discovery contract
• Code maturity assessment
• Contributing
• Support

Related repositories

• Hookbuilder builds, checks, and prepares applications.
• Programmable contains the platform, contracts, and Explorer.

Independence

Programmable Apply is independent open-source software. It does not claim affiliation with or endorsement by Uniswap Labs or the Uniswap Foundation.

Released under the MIT License.