Skip to content

Latest commit

 

History

History
68 lines (52 loc) · 4.36 KB

File metadata and controls

68 lines (52 loc) · 4.36 KB

Codex for Open Source form draft

Status: ready for applicant completion; not submitted and not an approval claim Last reviewed: 2026-08-23 Official form: https://openai.com/form/codex-for-oss/

This is a copy-ready draft for the current OpenAI form. The form asks for the applicant's ChatGPT email, public GitHub identity, maintainer role, public repository URL, OpenAI Organization ID, a qualification statement and a description of API-credit use. The qualification and credit-use fields have a 500-character limit; npm run check:application-dossier verifies them.

Applicant fields to complete manually

Applicant name: [FILL BEFORE SUBMISSION]
ChatGPT account email: [FILL BEFORE SUBMISSION]
Public GitHub username: daichunghy
Public repository URL: https://github.com/daichunghy/patchgate
Role: [CONFIRM PRIMARY OR CORE MAINTAINER]
Interest: [CONFIRM CODEX SECURITY, API CREDITS, OR BOTH]
OpenAI Organization ID: [FILL BEFORE SUBMISSION]

Copy-ready form answers

Why does this repository qualify? (maximum 500 characters)

PatchGate is an Apache-2.0 review-readiness gate: trusted policy, commit-bound checks, ownership and human boundaries before a PR consumes review time. Public beta Action, security model, documented shadow path. 1 GitHub star, 0 forks, no npm package (name taken), no external pilots (2026-08-23). Applying on ecosystem importance, not adoption.

How will you use API credits for your project? (maximum 500 characters)

API credits would support bounded issue triage, deterministic regression-fixture drafting, pull-request review preparation, authorized security-finding triage, and release/rollback checklists. Outputs would be tied to public issues, tests, PRs or release artifacts. The evaluator would remain deterministic; maintainers would approve code, security disclosure, pilots and releases; prompts would exclude secrets and private repository data.

Anything else we should know? (maximum 500 characters)

PatchGate addresses review burden when contributions move faster than maintainer attention. It does not detect AI authorship or replace human approval. Checkable evidence: public Apache-2.0 repo, v0.1.0-beta.5 shadow tag, CI on main, SECURITY.md. Discussions and issues are self-authored maintenance, not community adoption. No external pilots.

Evidence snapshot for the application

Signal Verified state on 2026-08-22 Evidence boundary
Repository visibility daichunghy/patchgate is public, Apache-2.0, default branch main Public foundation, not adoption
Default branch Current main@34d998b; hardening PR #9, follow-ups and current beta5 release merged by the maintainer (administrator decisions, recorded as such) Merged workflow; those merges lack independent approving review
Release Current tag v0.1.0-beta.5 at 34d998b; earlier beta tags are superseded Beta for shadow evaluation; not production-declared, not externally piloted
Dependency hygiene Dependabot PRs #11/#13/#14 merged after local re-verification; #12 (TypeScript 7) deferred with documented @vercel/ncc blocker Active maintenance, not a quality guarantee
Community activity Discussions #1, #2, #3, #8 and #10; community scheduler active on main Self-authored maintenance activity; no external replies yet
Usage signals No verified downloads, downstream users or pilots; check live star/fork counts at submission time Do not claim broad adoption
Pilot/release Shadow-installation no-go decision recorded in the release record; no completed external pilot Required evidence for any v0.1 claim remains open

Five-day completion checklist

  • Fill applicant name, ChatGPT email and OpenAI Organization ID.
  • Confirm primary/core maintainer role and public GitHub profile visibility.
  • Record any real external reply, pilot consent or contribution; do not substitute self-authored activity.
  • Re-run the live metrics check immediately before submitting.
  • Submit manually through the official form and retain the confirmation.

Selection is not guaranteed. The project should be described as a public, security-conscious pre-release project with a concrete ecosystem problem, not as widely used software.