Prompt-Case-Hub

A unified, format-standardized hub that aggregates prompt cases from popular open-source prompt repositories — image, video, and text — and re-expresses them under one consistent case schema.

Most prompt repos in the wild each invent their own structure: some are flat README tables, some are per-folder markdown, some are JSON, some bury the prompt inside prose. That makes them hard to consume programmatically and hard to compare. Prompt-Case-Hub does two things:

1. Resource integration — curates and credits high-quality cases from the open-source repositories listed below.
2. Format unification — normalizes every case into a single case schema (YAML front-matter + prompt body, plus a JSON index) so any tool can load, filter, and render cases the same way.

Curated by BlockRun.ai. This repo holds the format, index, and attribution layer — original prompt authorship and licensing stay with the upstream sources (see SOURCES.md).

---

Integrated sources

At minimum, the hub integrates and credits the following open-source repositories (full attribution + license notes in SOURCES.md):

Currently integrated — 848 cases normalized from these upstream repositories (via the seed aggregation repo BlockRunAI/Claude-Code-GPT-IMAGE2-SeeDance-BlockRun):

Cases	Repository	Focus
352	freestylefly/awesome-gpt-image-2	gpt-image-2 image cases
300	EvoLinkAI/awesome-gpt-image-2-prompts	gpt-image-2 prompt collection
105	YouMind-OpenLab/awesome-seedance-2-prompts	Seedance 2 video prompts
63	EvoLinkAI/awesome-seedance-2-guide	Seedance 2 guide + cases
28	ZeroLu/awesome-seedance	Seedance video prompts

Full attribution and candidate (not-yet-integrated) sources are in SOURCES.md. The list is additive — new sources are welcomed via PR as long as each integrated case carries its source attribution (see FORMAT.md).

---

Repository layout

.
├── FORMAT.md            # the unified prompt-case schema (the spec)
├── SOURCES.md           # integrated repos: attribution + license notes
├── schema/
│   └── case.schema.json # JSON Schema for validating a case's front-matter
├── cases/
│   ├── index.json       # machine-readable index of all cases
│   └── *.md             # one file per case (front-matter + prompt body)
└── CONTRIBUTING.md      # how to add / normalize a case

Using the cases

Every case is a markdown file with YAML front-matter and a fenced prompt body, plus an entry in cases/index.json. To consume them:

• Browse: open any file under cases/.
• Programmatic: read cases/index.json for the catalog, then load each case file and parse its front-matter against schema/case.schema.json.

See FORMAT.md for the full field reference and a worked example.

Status

848 cases normalized and live under cases/, indexed in cases/index.json. The schema, index format, and attribution layer are in place. Next up: integrating the candidate text/general sources listed in SOURCES.md. Contributions that add a correctly-formatted, properly-attributed case are welcome — see CONTRIBUTING.md.

License

The curation layer (schema, index, docs) is released under the MIT License (see LICENSE). Individual prompt cases retain the license and authorship of their upstream source — consult SOURCES.md before redistributing any case content.