rampstackco/claude-skills / skills/oss-contribution-research
oss-contribution-research
VerifiedResearch and prepare an outside contribution to an open-source repo, then stop at a queue item for a human to approve: policy check, intake, duplicate check, reproduction, patch, the repo's own checks, an independent critic rerun, and a disclosed, hashed draft of the exact text to post. Use this skill whenever the user wants to contribute to someone else's repo, fix an upstream bug, pick up a good-first-issue, draft an AI-assisted pull request for an outside project, follow up on their own outside PR, or run a gated contribution pipeline. Triggers on open source contribution, contribute to this repo, upstream fix, outside PR, external PR, good first issue, fix this issue in their repo, contribution research, AI-assisted PR, disclose AI in PR, reply to review on my PR. Also triggers when the user names an issue in a repo they do not maintain and asks for a fix, even if they do not say contribute.
| File | Kind | Size |
|---|---|---|
| README.md | markdown | 366 B |
| SKILL.md | skill md | 18.8 KB |
| references/example-intake-item.md | markdown | 3.3 KB |
| references/policy-classification.md | markdown | 5.0 KB |
| references/pr-description-template.md | markdown | 5.3 KB |
| references/queue-item-template.md | markdown | 6.6 KB |
| references/repro-record-template.md | markdown | 4.9 KB |
references/policy-classification.md
View on GitHub# Policy classification
How to find a repo's AI-assisted contribution policy, classify it, cite it, and decide eligibility. Run this at gate 2. The registry may already hold a classification. Re-check the cited file anyway if it is older than a month, because policies change.
## Where to look, in order
1. **Contributor guides.** Check `CONTRIBUTING.md` at the root, in `.github/`, and in `docs/`.
2. **Org-wide files.**
- The org's `.github` repo, which supplies a default `CONTRIBUTING.md` when the repo has none.
- Org policy files such as an `AI_POLICY.md` that say they apply to every repo in the org.
3. **The pull request template.** Check `.github/pull_request_template.md` and `.github/PULL_REQUEST_TEMPLATE/`.
4. **Other `.github/` files.** Issue templates, and workflows that auto-close pull requests. Search workflow files for `close` and for author-association checks.
5. **Pinned issues.** Policy announcements often live there.
6. **Agent guides.** Check `AGENTS.md` and `CLAUDE.md`. These often address agents working inside the repo, not contributors, so read them as two kinds of text:
- Lines aimed at contributors' agents, such as disclosure rules or rejection warnings, count as policy.
- Lines aimed at the maintainers' own agents do not count as policy, but they can still constrain disclosure. One example is a ban on attribution footers.
## Classes
| Class | Test | Result |
|---|---|---|
| `allows` | An explicit line welcomes AI-assisted contributions, usually with conditions | Eligible, subject to its conditions |
| `silent` | Nothing found in any location above | Eligible; pull request descriptions carry the default prose line, comments and replies carry none |
| `bans` | An explicit line refuses AI-generated or AI-assisted contributions | Stop |
Cite the exact line with the file and line number, for example `CONTRIBUTING.md:63 "AI-assisted contributions are welcome"`. Or write `none found` and list the locations checked. A paraphrase is not a citation.
## Eligibility stops beyond `bans`
Stop even when the class is `allows` or `silent` if any of these hold. Each is an exclusion. A fuller review flow does not lift it, and the agent never fills in a human attestation on the user's behalf.
- **Attestation.** The policy requires the human owner to attest to reviewing the complete change.
- **Issues only.** The repo accepts issues but not pull requests, for example "We accept issues, not pull requests."
- **Members only.** The repo accepts pull requests only from members or staff. A workflow that auto-closes outside pull requests counts.
- **Assignment first.** Outside pull requests are reviewed only after a maintainer assigns the linked issue. This is a stop until the assignment arrives, and the intake gate handles it.
Data can also exclude a repo when no line says so. If the last 20 outside pull requests got no maintainer response and none merged, the repo is not absorbing outside work. That is the registry owner's call, recorded as status `excluded`.
## Disclosure patterns seen in practice
Record the pattern per action type in the registry, as a prose rule plus a template the poster checks: `unknown`, `none`, or one line with `{slots}`. The skill reads the template from the registry and does not guess.
- **Disclosure with extent.** "This must be disclosed in the pull request or issue, along with the extent to which AI assistance was used." Often paired with "If your PR responses or comments are being generated by an AI, disclose that as well," which covers replies too.
- **Fixed description line.** A line in which the agent names itself and links to itself, plus self-identification as automated in every comment and review. Some of these repos also require updating an existing agent comment instead of posting another. The poster only posts new comments, so such a repo stays paused.
- **Optional disclosure.** The policy welcomes AI assistance and makes disclosure optional. Record the line that says so; that is not `none found`. The registry owner picks the template, usually the same prose line as a silent repo.
- **Environment disclosure.** The model, harness, harness version, and every installed plugin, on every pull request and issue.
- **No AI attribution.** An agent guide forbids attributing commits or pull request bodies to an AI tool. Disclosure is one prose line in the body with no tool or model name, and nothing in the commits.
When a repo's lines cover pull requests but say nothing about comments, the comment format is `unknown`. An `unknown` format stops that action type until the registry owner fills it in. This applies only to repos with an explicit policy. A `silent` repo has no lines to follow, so it uses the defaults in `pr-description-template.md`: one prose line in pull request descriptions, and none on comments and replies.
## Recording the result
Give one line per repo in the Policy section of the queue item:
```
allows: CONTRIBUTING.md:63 "AI-assisted contributions are welcome" (https://...). Disclosure for issue_comment: [registry text].
```
Related skills
| Skill | Stars | Updated | Works with | Status | Install |
|---|---|---|---|---|---|
| writing-skillsobra/superpowersUse when creating new skills, editing existing skills, or verifying skills work before deployment | Stars: 322k | Updated | Claude CodeCodex | Verified | Install |
| xlsxanthropics/skillsUse this skill any time a spreadsheet file is the primary input or output. This means any task where the user wants to: open, read, edit, or fix an existing .xlsx, .xlsm, .xltx, .csv, or .tsv file (e.g., adding columns, computing formulas, formatting, charting, cleaning messy data); create a new spreadsheet from scratch or from other data sources; or convert between tabular file formats. Trigger especially when the user references a spreadsheet file by name or path — even casually (like "the xlsx in my downloads") — and wants something done to it or produced from it. Also trigger for cleaning or restructuring messy tabular data files (malformed rows, misplaced headers, junk data) into proper spreadsheets. The deliverable must be a spreadsheet file. Do NOT trigger when the primary deliverable is a Word document, HTML report, standalone Python script, database pipeline, or Google Sheets API integration, even if tabular data is involved. | Stars: 198k | Updated | Claude CodeCodex | Verified | Install |
| shipping-and-launchaddyosmani/agent-skillsPrepares production launches. Use when preparing to deploy to production, or when asking what needs to be in place before shipping. Use when you need a pre-launch checklist, when setting up monitoring, when planning a staged rollout, or when you need a rollback strategy. | Stars: 119k | Updated | Claude CodeCodex | Verified | Install |
| childhoodthedotmack/claude-memBabysit a shipped release after merge, publication, or deployment. Verify published packages, production behavior, installed plugins, and late CI or user reports until the observation window is clean; use babysit for PR readiness before release. | Stars: 118k | Updated | Claude CodeCodex | Verified | Install |