How to share solved problems with Unknown
The author tested Shared Knowledge MCP through scripts and GitHub Copilot Agent mode, producing reviewed Markdown contributions and documentation with generated audio.
dev.toPrompt to try
Editorial adaptation, not the source author’s exact prompt. Replace bracketed inputs and review tool permissions first.
Using Shared Knowledge MCP, search for existing coverage of [solved technical problem]. Draft a standalone English Markdown article containing only [approved solution details], validate it, and open a pull request for human review. Do not publish the private conversation.
Reach people exploring AI agents
See placements →What happened
The author reports a live documentation site with four published articles and audio versions, and tests through both script execution and Copilot Agent mode.
Reported by pascal_cescato_692b7a8a20 on dev.toHow the agent did it
- Convert a solved dependency problem into structured Markdown.
- Validate the contribution and open a GitHub pull request.
- Address Copilot feedback on formatting and headings.
Where a person stepped in
Sharing requires an explicit user choice, and publication follows human review and merge; the first contribution also received automated Copilot feedback.
A person approved each step before the agent continued.
Shared Knowledge MCP was built to turn a solved technical problem into a reusable article without publishing the underlying private conversation. The author describes a pipeline in which a user explicitly chooses to share, the solution becomes structured English Markdown, and the MCP server validates it and opens a GitHub pull request. Human review and merge precede public documentation and generated audio.
The first reported contribution covered an optional Python dependency that crashed an import chain. It became an article with sections, metadata, and tags. GitHub Copilot reviewed the pull request and flagged YAML formatting and heading issues. Fixes were applied before merge, after which an ElevenLabs audio version appeared on the documentation site.
The author then tested a conversational route using GitHub Copilot Chat in Agent mode inside a Codespace. In that session, the assistant first searched the existing knowledge base for duplicates before deciding to publish. The author says this choice was made by the assistant rather than hard-coded in the server. The documentation site held four published articles with audio versions. Who it's for: developers who want to convert resolved technical investigations into reviewed community documentation while choosing explicitly what to share.
Prompt
Suggested prompt (written by Agents In Use, not the author's)
Source: dev.to · pascal_cescato_692b7a8a20
Frequently asked questions
Does the system publish conversations automatically?
No. The user explicitly chooses to share the relevant solution, and publication follows human review and merge of a GitHub pull request.
What did Copilot do in the conversational test?
GitHub Copilot Chat in Agent mode queried search_knowledge to check for duplicates before deciding to publish.
More details
Remaining steps
- Merge the contribution and generate its audio version.
- Test Copilot Agent mode searching for duplicates before publishing.
About this account
Copilot Agent mode independently searched the knowledge base before publishing; the MCP server validated contributions and opened pull requests.
Platform: dev.to. Attribution: https://dev.to/pascal_cescato_692b7a8a20/from-ai-solutions-to-shared-knowledge-building-an-mcp-for-the-community-6bk. Prompt: Suggested (written by Agents In Use, not the author's). Sourced by the Agents In Use pipeline.
Credit: pascal_cescato_692b7a8a20 on dev.to
Source published 7 Sept 2026
Source checked 14 Oct 2026
Tools and topics
GitHub Copilot Chat · Shared Knowledge MCP · GitHub Codespaces · ElevenLabs · Astro · Starlight
MCP · Knowledge sharing
Editorial completeness: 100/100. Measures documented fields, not agent performance or verified outcomes.