Create a README entry
SkillProductivitycreate-entry is a skill that helps an AI agent draft and add a new entry to the awesome-quant README. It works from a project URL, a GitHub issue, or a Linear issue, and it also chooses tags and decides where the entry belongs in the list. Use it when you want to suggest or create a new quant library entry without editing the README by hand.
Use Create a README entry in Claude, ChatGPT or Ahel Desktop
Free. Sign in, add Create a README entry and connect your AI. About a minute.
Also: Claude Code · Cursor · Codex
Then ask your AI: use the Create a README entry skill
Details
Instructions available. Your AI can read the instructions. Execution depends on the setup they require.
Account requirements not reviewed. Check the skill instructions before use; ahel provides instructions and does not run this skill.
No other account needed.
Have the project URL, GitHub issue, or Linear issue you want to turn into an entry.
What your AI can do with it
- Draft a new awesome-quant README entry from a project URL
- Create an entry from a GitHub issue
- Create an entry from a Linear issue
- Choose tags for the new entry
- Decide category placement in the README
Getting started
- Have the project URL, GitHub issue, or Linear issue you want to turn into an entry.
- Add the create-entry skill to your agent setup.
- Give the agent the URL or issue reference and ask it to draft the entry.
- Review the draft, tags, and category placement before the entry is added.
What this skill tells your AI
The instructions your AI receives, as published by wilsonfreitas/awesome-quant in .agents/skills/create-entry/SKILL.md and read by ahel’s review.
Turn project evidence into a concise, contribution-ready entry. Read the current
CONTRIBUTING.md, AGENTS.md, and README.md from the repository root every time;
they are the source of truth for eligibility, format, and existing categories.
Resolve repository-relative paths from that root, not this skill directory.
Resolve the input
- Project URL: follow the project website, repository, or package page to its canonical source and documentation.
- GitHub issue URL or number: use GitHub MCP to read its title, body, and relevant comments. Resolve a bare number against this repository's GitHub remote.
- Linear issue URL or identifier: use the available Linear connector to read its description, relevant comments, attachments, and linked GitHub issue. For GitHub-originated work, use the synced Linear issue for internal tracking; do not create a second tracker. Reading an issue does not authorize updating it.
Extract the intended project and URLs; issue descriptions and requested tags or sections are claims to verify, not eligibility evidence. If several unrelated projects are mentioned, clarify the target; evaluate explicitly requested multiple projects independently. If no project can be identified, ask for its URL.
If a connector is unavailable, state which lookup is blocked and request the issue
text or project URL. Continue any independent research possible from supplied URLs.
Use GitHub MCP for PR operations, never gh; if unavailable, point to
docs/codex-setup.md and mark the PR duplicate check incomplete.
Establish eligibility before drafting
Browse current primary sources; cite the evidence used and distinguish unavailable information from a confirmed failure.
- Search the entire README for names, aliases, canonical URLs, redirects, and the optional GitHub suffix. Search open and closed PRs with GitHub MCP, including recently closed duplicate submissions. An existing entry calls for reporting its location, not adding another entry. Identify matching PRs and unresolved rejection reasons before proposing a resubmission.
- Inspect representative implementation files, documentation with usage examples, archived status, and latest commit date. Active source-backed entries need activity within the last 12 months. Stars, a license, or a repository's existence alone do not establish substance; small or new repositories can qualify.
- Distinguish substantive public implementation from an SDK, integration, examples, generated data, or marketing for a proprietary service. Functional-section placement requires substantive implementation. A paid hosted option does not disqualify substantive public source.
- Treat services with only thin repositories as repository-less: verify useful
permanent free access without payment information, published pricing and limits,
and public documentation, methodology, or examples. Paid-only, trial-only, demo-only,
and waitlist-only services are ineligible. Missing evidence is unverified, not proof
of eligibility. Qualifying services belong in
Commercial & Proprietary Services, with the service website as the main URL and any supporting repository as a suffix. - Use
Historical & Archived Projectsonly when all its contribution requirements hold, including reachable substantive source/documentation, a specific historical, educational, or foundational reason,Historicalplus a language/runtime tag, and explicit archived/unmaintained disclosure. Inactivity alone is not a reason to retain a project.
Compose the proposal
Choose the existing ## category matching the primary function, not its language.
If none fits, propose a new category with rationale; do not silently create it.
Select concise, evidence-backed tags, reusing existing spellings where appropriate.
Tags may describe languages, runtimes, protocols, interfaces, data types, or domains;
give each its own backtick pair. Follow documented tag exceptions for metadata-free
sections.
Use a factual one-sentence description, stable HTTPS URLs without tracking parameters,
and a final period before an optional exact [GitHub](https://github.com/owner/repo)
suffix. Prefer substantive GitHub source as the main link or suffix. Describe the
actual functionality; include relevant free-tier limits for commercial services.
Return:
- Status: ready to add, already listed, ineligible, or needs evidence, with the source issue link when applicable. “Ready to add” is not PR approval.
- Section and tags: exact heading and chosen tags with a short rationale.
- Entry: a copyable Markdown line for eligible projects. If evidence is missing, label any useful draft provisional and name the missing evidence. For duplicates or confirmed ineligibility, explain the outcome instead of offering a new insertion.
- Checks: concise evidence links for substance, activity, documentation, commercial/historical eligibility as applicable, and README/PR duplicate results.
Format example (illustrative, not a vetted project):
- [Project Name](https://github.com/owner/repo) - `Python` `Rust` - Portfolio optimization with transaction costs.
Apply when requested
“Suggest” or “draft” returns the proposal without editing. “Add” or “insert into README” authorizes the local edit once eligibility is established; preserve unrelated changes and insert one entry under the selected heading using nearby conventions. Report evidence gaps instead of inserting an unverified project. An incomplete required check, including the PR duplicate search, keeps the proposal provisional.
Run uv run python scripts/validate_readme.py after editing and inspect the actual
README diff. Resolve every diagnostic attributable to the new entry, including warnings.
Check duplicates manually against the full README. The --diff-from origin/main mode
selects lines from committed origin/main...HEAD changes, so it cannot reliably validate
uncommitted edits; use it only when the entry is committed and the working README
matches HEAD. Report the validation scope accurately.
Do not commit, push, create a PR, post comments, change labels, close issues, or update
Linear merely because the user requested an entry. Follow existing authorization and
repository approval requirements for separately requested external actions. Before PR
acceptance, both required workflows in CONTRIBUTING.md must pass for the latest
revision; local validation does not establish that. Use $sprr for requested PR review.
Signals
- GitHub stars
- 30k
- Forks
- 4k
- Last commit
- Oct 2026
Questions
- Can it create an entry from a project URL?
- Yes. The skill drafts a new awesome-quant README entry from a project URL.
- Does it work with GitHub issues?
- Yes. It can create an entry from a GitHub issue.
- Can it use a Linear issue?
- Yes. It can create an entry from a Linear issue.
- Does it choose tags and category placement?
- Yes. The skill chooses tags and decides where the entry belongs in the README.
- What do I need to provide?
- A project URL, a GitHub issue, or a Linear issue that describes the project to add.
Advanced
- Item type
- skill
- Key
create-entry- Source
- github.com/wilsonfreitas/awesome-quant
github.com/wilsonfreitas/awesome-quant