friuns2/BlackFriday-GPTs-Prompts: a markdown index of free GPTs and jailbreak prompts

What the BlackFriday-GPTs-Prompts repository actually contains

The repository is a flat set of markdown files. The top level holds a README, a LICENSE, a Models.md, and one file per topic: Programming.md, Marketing.md, Academic.md, Job-Hunting.md, Game.md, Creative.md, Prompt-Engineering.md, Business.md, Productivity-&-life-style.md, Jailbreaks.md, and a dated 2024-May-Update.md. There is also a gpts/ directory. The README lists each of these as a link, which is the entire navigation model: you open the file for the category you care about and read it.

The stated purpose is in the repository description: a list of free GPTs that does not require a Plus subscription. That is a narrower claim than it first appears. It is about access, not about capability or quality. The README does not explain how the access works, whether the links are custom GPT share links, whether they are third-party mirrors, or what happens when a link stops working. Nothing in the README documents a verification date per entry.

The audience is therefore anyone who wants prompt text and GPT links as raw material: someone drafting a prompt library, someone curious what other people put in these files, someone collecting examples for a writeup. It is not a library you import and it is not a service you call.

Why there is nothing to install and what that means for reuse

There is no package, no binary, and no runtime. The repository is consumed by reading it, and the practical workflow is a browser tab. The README gives no installation section because there is nothing to install.

If you want the files locally so you can search across all of them at once, the README points to the repository on GitHub, which is the address you would clone from. The README itself does not show a clone command, so the only setup step it documents is opening the repository page and reading the files there.

What you get either way is the same set of category files, and searching them locally is more practical than clicking through the GitHub web UI when you are looking for a phrase across every file. The README does not describe any other supported way to consume the content, and there is no generated site, index, or feed mentioned for this repository.

The trade-off of a pure markdown collection is that nothing enforces structure. Entries are whatever the contributors typed. The README does not state a required format for a new entry, so a pull request adding a link has no schema to satisfy and no test that would fail if the link is dead.

The AnyClaw banner at the top of the README

The first thing in the README is not about prompts. It is a centred block advertising AnyClaw, described as OpenClaw plus Codex running natively on Android, in one APK, with no root and no Termux, linking to a separate repository and to a download page on the project's GitHub Pages site. Two screenshots are embedded from the raw content host.

This matters for how you read the repository. The README's own homepage field points at that Android assistant page, not at a documentation site for the prompt collection. A reader arriving from a search result may reasonably expect a prompt index and find a product banner first. The README does not explain the relationship between the two projects beyond the shared author, and it does not say whether the prompt files are maintained alongside the Android work or frozen while attention moved elsewhere.

There is also a plain URL to an app-ads.txt path under the same Pages domain. The README does not describe what it is for. If you are evaluating this as a reference collection, treat the top of the README as advertising and start reading from the file list below it.

Jailbreaks.md and the gap between the MIT licence and the content

The repository carries an MIT licence, and the LICENSE file is at the top level. MIT is a permissive software licence: it grants rights to use, copy, modify, and distribute the covered work with the copyright notice and permission notice retained, and it disclaims warranty. Applied to a repository of markdown text, that is a reasonable fit for the prose the contributors wrote themselves.

Jailbreaks.md is the file that makes the licence question less tidy. Prompt text designed to get a model to bypass its own restrictions is not the same kind of artifact as a hello-world script, and the MIT grant does not resolve what happens when you paste that text into a hosted service. The terms of that service govern the interaction, not the repository's licence. The README does not discuss this tension at all, and it does not flag the jailbreak file as distinct from the rest. That silence is the main thing to weigh before reusing content from that file in anything you publish.

This is not legal advice, and the repository's own LICENSE file is the document that matters. Read it rather than the badge.

Maintenance signals: last push on 2026-03-18 and no releases

The repository is not archived, and the last push was on 2026-03-18. That is the only maintenance fact available here. No releases were retrieved, so there is no versioned artifact to track and no changelog to read. The most recent dated file in the tree is 2024-May-Update.md, which is a snapshot of one month's additions rather than a running log.

For a prompt collection this is a different kind of risk than for a library. Nothing breaks when it goes stale, because there is no interface to break. What ages is the content. GPT links rot, models change behaviour, and prompt text that produced a given result at one point may not produce it later. The repository has no mechanism visible in the README for marking an entry as dead or superseded, and no CI that could check links.

Upgrade cost is correspondingly low and correspondingly manual: pull the branch and diff your local copy against it. There is no migration path to plan because there is no dependency to migrate. The cost you actually pay is the time spent checking whether the entries you care about still point somewhere useful.

How it compares to prompt libraries you can import

The obvious alternative is a prompt library that ships as a package, where prompts are versioned, typed, and loadable from code, so a test can assert that a prompt still exists and a lockfile pins the revision you depend on. That approach solves a problem this repository does not attempt to solve: reproducibility. If you need the same prompt string in a build six months from now, a markdown file in a git repository gives you that only as long as you pin the commit yourself.

The difference in approach is structural. An importable library treats a prompt as an artifact with an owner and a release. This repository treats prompts as reading material organised by human category names, and its value is breadth and the low effort of browsing. The README's organisation by use case, programming, marketing, academic, job hunting, and so on, is the feature. You will not find that taxonomy in a package's type definitions.

A second alternative is simply keeping your own notes. That is what this repository effectively is for its maintainers, and copying the parts you need into your own file removes the dependency on someone else's link hygiene entirely.

Who should clone it and who should keep looking

Clone it if you want prompt text to read, adapt, and file into your own system, and if you are treating every entry as a starting point rather than a verified tool. The category split means you can pull one file, Programming.md or Academic.md, and ignore the rest, which keeps the surface small.

Do not clone it if you need a component in a pipeline. There is no API, no export format documented in the README, and no release to pin. Do not rely on it as a citation for what a particular GPT does, because the README records no per-entry date and the last push on 2026-03-18 tells you when the repository changed, not when any individual link was last confirmed working.

The first thing to verify is the file you actually intend to use. Open the category file, check that the links resolve, and decide whether the entries still match your model. The second thing is the LICENSE file, particularly if your use touches Jailbreaks.md.

Editorial conclusion

Adopt this repository if you want a browsable set of prompt text and GPT links you can copy into your own notes, and if you are comfortable that the README's central claim about Plus-free access is unverified and that the jailbreak file sits in the same tree under an MIT licence. Do not adopt it if you need a maintained tool, a stable API, or a source you can cite for what a GPT does today, because the only maintenance signal is a last push on 2026-03-18 and the README records no releases. Verify first whether the links in the category file you care about still resolve, and read the LICENSE file rather than assuming MIT covers every file's contents.

Frequently asked questions

Official sources

1. friuns2/BlackFriday-GPTs-Prompts on GitHub
2. Issues
3. License: MIT
4. Project website
5. README