io.github.ADahlingPublished by GitHub account ADahlingv1.0.0Updated Oct 25, 2026
Search 3,323 graded AI prompts for software builders; fetch any prompt by slug. No key needed.
A public, keyless MCP server that lets an assistant search and fetch 3,323 rubric-graded AI prompts for software delivery across six stages.
Web executable via Streamable HTTP. Remote servers run from the web runtime once configured in a workspace.
Paste this prompt into Claude Code, Codex, Cursor or another agent — or into SourceWeft chat. The agent reads this server's install guide, shows you its publisher, endpoint and sign-in, and adds it to your client once you agree. Keys never go into a command.
Read https://sourceweft.com/mcp/io.github.ADahling%2Fbuilders-prompt-library/install.md and connect the MCP server it describes to my client. Before changing any config, show me its publisher, endpoint and sign-in method and wait for my OK. Never put an API key in a command or in a file I would commit; ask me where it should come from.Add this to your client's mcpServers config.
{
"mcpServers": {
"the-builder-s-prompt-library": {
"type": "http",
"url": "https://buildersprompt.com/mcp"
}
}
}3,323 rubric-graded AI prompts for every stage of shipping software, from idea to launch. This is the open data behind buildersprompt.com, exported from the live catalog so you can grep it, vendor it, load it into your own tools, or point an agent at it.
Built by Alicia Dahling and Peter Lowe. Free to read, copy and adapt under CC BY 4.0 (see LICENSE): keep the attribution, do what you like with the prompts.
Fill-in fields look like {{project_name}}. Replace them before you paste.
Every prompt is scored 0 to 100 on a five-part writing rubric (clarity, specificity, format, variables, actionability) and stamped A+ (90+), A (80 to 89), B (70 to 79) or C (60 to 69). Anything below 60 is not published. The grade describes how well the prompt is written, not how well a model answered it. Prompts that were actually run in ChatGPT, Claude and Gemini, with the graded answers, are shown on their pages at buildersprompt.com under "Tested in real AI models". Twenty-two prompts were read line by line by a human and carry hand_reviewed: true.
The library is also a public MCP server. Point Claude, Cursor, ChatGPT or any MCP client at https://buildersprompt.com/mcp (Streamable HTTP, JSON-RPC 2.0, no key) and you get three read tools: search_prompts, get_prompt and list_stages.
This export is regenerated from the live database. Each prompt's updated_at is the last edit on the site. If you need a specific prompt's current text, its url is the source of truth.
Please credit "The Builder's Prompt Library, buildersprompt.com" and link back when you republish prompts from this set. That is the whole license.
Source: README.md at commit e9e31c7
Search prompts
Search the Builder's Prompt Library (3,323 rubric-graded prompts for software delivery) by keyword. Optionally filter by delivery stage or prompt-writing grade. Returns ranked matches with slug, title, summary, stage, category, and grade. Grades and quality_score describe prompt writing, not tested model performance.
gradelimitqueryrequiredstageGet prompt
Fetch the full text and metadata of a single published prompt by its slug, including where it runs (any chat app, chat app with your code pasted in, a tool with web access, or a coding agent).
slugrequiredList stages
List the six delivery stages in the Builder's Prompt Library and how many published prompts each contains.
Takes no parameters.
90% of the applicable checks met
How complete this listing is, computed from the registry entry, the endpoint and public sources. Not a rating of the server and not a statement that it is safe or working today: reachability is a dated check, and no installs or reviews go into the score.
Declare a license in the repository or the registry entry.