Sanjay C./Contribution Infrastructure
Independent research · Bengaluru

Open source contribution infrastructure

Sanjay C.

Evidence and small tools for parts of open source that never got the infrastructure code review did — localization and language inclusion, accessibility review, AI contribution policy, and repositories legible to the AI agents that now read them.

See the work GitHub

Current work

One method across four domains: document how the gap shows up in real repositories, then ship the smallest useful artifact.

  • Shippable tool

    Localization & language inclusion

    Most open source ships English-first. The bottleneck on Indic locales is rarely translation quality — it is review and operations.

    oss-language-inclusion · i18n-security-lint on PyPI

    Evidence on how localization pull requests actually get reviewed upstream, and what makes maintainers merge them — plus the tooling that gap needs. Includes i18n-security-lint, a CI linter for four defect classes in translated strings: bidi overrides, XSS fragments, format-string drift, interpolation breakage. Prior writing in CACM and DevOps.com.

    Open the repo pip install i18n-security-lint PyPI
  • Active research

    Accessibility contribution review

    We scored seven real accessibility PRs against a 12-point rubric. None reached the mature band, and the most heavily reviewed one tied for the lowest score.

    oss-accessibility-inclusion · rubric + case studies + templates

    A meta-review of how accessibility pull requests are actually reviewed — plus reusable ACCESSIBILITY.md, issue, and PR templates so the next fix is reviewable by a non-expert.

    Open the repo Steal the template Review rubric
  • Shippable tool

    Machine-operable documentation

    The examples in your README are the most-copied, least-tested code you ship — now to AI agents as much as to people.

    machine-operable/readme-ci · extractor + sandbox runner on PyPI

    readme-ci finds the code examples in Markdown docs and runs them in a locked-down sandbox, so “our examples still work” becomes a CI check instead of a hope. First tool of Machine-Operable Open Source — a program on repositories that machines, not only people, now read.

    Open the repo pip install readme-ci PyPI
  • Draft standard

    AI contribution policy

    Nine projects. Five incompatible postures. One proposed machine-readable shape.

    oss-ai-contribution-policy · catalogue + schema v0.1

    A verified catalogue of what curl, Ghostty, LLVM, tldraw, and peers actually wrote — and a draft ai-contribution-policy.yml built on verification over detection, not bans.

    Open the repo Policy catalogue Copy a YAML example

Method

Code contribution got linters, CI, and CODEOWNERS. Language work, accessibility work, and AI-assisted work mostly did not. I study that asymmetry in public repositories, then publish the smallest fix the evidence supports.

  1. Study real projects and primary sources.
  2. Name the recurring infrastructure gap.
  3. Score or catalogue what communities already invent.
  4. Ship a small reference artifact (linter, rubric, schema).
  5. Invite critique from maintainers, practitioners, and foundations.

Portfolio label: OSS Infrastructure Initiative. The public face of the work is the named researcher and the artifact, not an abstract organization.

Writing

Background

  • 30+ years in capital markets technology and product/program management — including eleven years in the United States (UBS, Deutsche Bank, Point72).
  • Visiting faculty in management programs (strategy, finance, risk).
  • Independent advisor to early-stage founders on AI and product strategy.
  • Open to talks, briefings, and maintainer feedback on contribution infrastructure.

Contact

I welcome conversations with maintainers, OSPOs, foundations, researchers, and journalists working on these gaps.

GitHubgithub.com/ecogetaway
Mastodon@ecogetaway@mastodon.social
LinkedInlinkedin.com/in/sanjay-c-1464602
Emailecogetaway@gmail.com