# AI use cases for Peerbridge MCP (Hoylon/peerbridge-mcp) Repo: [Hoylon/peerbridge-mcp](https://github.com/Hoylon/peerbridge-mcp) | Use cases: 4 | Page: https://dedupedai.com/use-cases/hoylon-peerbridge-mcp/ PeerBridge MCP is a local-first, auditable control room where Codex, Claude Code, Grok, Kimi, provider APIs, and local models work as one governed team on code review and remote work. This file holds every use case of this repo. Each one starts with a "## " heading and ends with its URL. - Run the provider-free coordination demo: https://dedupedai.com/use-cases/hoylon-peerbridge-mcp/run-provider-free-coordination-demo/ - Run Implement and Review on one repo from the Windows app: https://dedupedai.com/use-cases/hoylon-peerbridge-mcp/run-implement-and-review-from-windows-app/ - Continue an earlier agent conversation in a writable room: https://dedupedai.com/use-cases/hoylon-peerbridge-mcp/continue-earlier-agent-conversation-in-new-room/ - Keep approved project decisions in a source-bound memory ledger: https://dedupedai.com/use-cases/hoylon-peerbridge-mcp/keep-approved-project-decisions-in-memory-ledger/ ## How to run the provider-free coordination demo with Peerbridge MCP Repo: [Hoylon/peerbridge-mcp](https://github.com/Hoylon/peerbridge-mcp) | Difficulty: Intermediate | Time: 30 minutes | Basis: From the README Run the synthetic demo and check the receipt for writer lease, quorum, and audit results. **Who it is for:** A maintainer or evaluator who wants proof that the coordination core blocks overlapping writers before trusting it with real agents. **The problem:** Claims about write leases, peer review quorum, and audit chains are hard to judge from a README alone. **The outcome:** A public receipt showing an overlapping second writer was rejected, two reviewers met quorum, completion rehashed the artifact, and the audit chain verified. ### Steps 1. `git clone https://github.com/Hoylon/peerbridge-mcp` (inferred, check the README) 2. `cd peerbridge-mcp` (inferred, check the README) 3. `python examples\demo_workflow.py --workspace demo-workspace --scope demo` 4. Read the printed receipt and confirm the overlapping writer was rejected and the audit chain verified with zero writes. ### Prompt to paste into your agent ```text Clone the Hoylon/peerbridge-mcp repo and run examples\demo_workflow.py with --workspace demo-workspace and --scope demo. Report the receipt, and verify that it shows the overlapping second writer rejected and the audit chain verified. ``` ### Caveat The project is alpha and the public API and database schema may change before 1.0, so treat the demo as evaluation, not a production guarantee. Disclaimer: Third-party repos we don't own or vouch for. AI-generated use cases can be wrong: review everything and run at your own risk. Full text: https://dedupedai.com/disclaimer/ URL: https://dedupedai.com/use-cases/hoylon-peerbridge-mcp/run-provider-free-coordination-demo/ ## How to run Implement and Review on one repo from the Windows app with Peerbridge MCP Repo: [Hoylon/peerbridge-mcp](https://github.com/Hoylon/peerbridge-mcp) | Difficulty: Intermediate | Time: 60 minutes | Basis: From the README Use the portable Windows build to have one agent write in a worktree while another reviews. **Who it is for:** A developer on Windows with Codex or Claude Code installed who wants a second agent to review changes before accepting them. **The problem:** Two agents editing the same repository overwrite each other and reviews often approve files that were never tested. **The outcome:** One agent writes inside an approved isolated Git worktree, a peer reviews the source, and the Trust Timeline shows whether the evidence is still current. ### Steps 1. Download `PeerBridgeControlRoom-0.1.0a6-windows-x64-portable.zip` from the Alpha 6 release page. 2. Compare the ZIP's SHA-256 with `SHA256SUMS.txt` on the release page. 3. Extract the complete ZIP to a writable folder and double-click `Launch PeerBridge.cmd`. 4. In the Modern Workbench, start the Implement + Review template on your local Git repository and approve one isolated Git worktree for the writer. 5. Open the Trust Timeline and confirm the review is bound to the current source before accepting completion. ### Prompt to paste into your agent ```text Using the Hoylon/peerbridge-mcp portable build, start an Implement + Review operation on with Codex as writer and Claude Code as reviewer. Report the worktree path and the Trust Timeline status, and verify that the review shows no stale-evidence warning. ``` ### Caveat The Windows portable build is an Alpha release, and it does not include provider credentials, so you must have your own agent accounts set up. Disclaimer: Third-party repos we don't own or vouch for. AI-generated use cases can be wrong: review everything and run at your own risk. Full text: https://dedupedai.com/disclaimer/ URL: https://dedupedai.com/use-cases/hoylon-peerbridge-mcp/run-implement-and-review-from-windows-app/ ## How to continue an earlier agent conversation in a writable room with Peerbridge MCP Repo: [Hoylon/peerbridge-mcp](https://github.com/Hoylon/peerbridge-mcp) | Difficulty: Intermediate | Time: 30 minutes | Basis: From the README Import a Codex or Claude Code history and continue it in a normal room without changing the source. **Who it is for:** A developer who had useful work in a past Codex or Claude Code session and wants to resume it inside a governed multi-agent room. **The problem:** Past agent sessions are scattered across tools and resuming them loses the context and the record of what was decided. **The outcome:** A new writable room that carries a bounded source-conversation ID and a SHA-bound context message, while the imported source room stays immutable. ### Steps 1. Open the Agent history import view in the Modern Workbench. 2. Review the bounded metadata list, which selects nothing by default. 3. Check only the conversations you want to import. 4. Select **Continue from history** on the imported conversation to create a writable room. 5. Confirm the new room shows the SHA-bound context message before you send the next task. ### Prompt to paste into your agent ```text In the PeerBridge MCP control room, import only the Codex conversation from the agent history list and use Continue from history to create a writable room. Report the new room ID, and verify that the source room remains read-only. ``` ### Caveat Hidden chain-of-thought and uncaptured external terminal history are not imported, and PeerBridge does not label them as visible. Risk, credentials: Import reads local session records from other agent tools, so check only the conversations you intend to bring into a room. Disclaimer: Third-party repos we don't own or vouch for. AI-generated use cases can be wrong: review everything and run at your own risk. Full text: https://dedupedai.com/disclaimer/ URL: https://dedupedai.com/use-cases/hoylon-peerbridge-mcp/continue-earlier-agent-conversation-in-new-room/ ## How to keep approved project decisions in a source-bound memory ledger with Peerbridge MCP Repo: [Hoylon/peerbridge-mcp](https://github.com/Hoylon/peerbridge-mcp) | Difficulty: Advanced | Time: 60 minutes | Basis: From the README Save room decisions as memory records that stay tied to their source evidence. **Who it is for:** A team lead running several agents over weeks who needs decisions to persist without silently merging unverified notes into project context. **The problem:** Agents forget earlier decisions, and copied notes lose the evidence that supported them. **The outcome:** Memory records at Private, Room, or Project scope, each SHA-bound to its explicit source evidence, with Project records requiring human approval. ### Steps 1. Open the Memory Ledger in the Modern Workbench. 2. Create a Room record from a decision that has explicit source evidence in the room history. 3. Review the record and its SHA binding before promoting it. 4. Approve the record as a Project record so later task briefings can use it. 5. Check that the record still lists its source evidence after a restart. ### Prompt to paste into your agent ```text In the Hoylon/peerbridge-mcp control room, create a Room memory record from the decision in and ask a human to approve it as a Project record. Report the record ID and its source SHA, and verify that the record links to the original evidence. ``` ### Caveat Memory records are only as reliable as the source evidence behind them, and the README does not describe how memory is shared outside the local machine. Disclaimer: Third-party repos we don't own or vouch for. AI-generated use cases can be wrong: review everything and run at your own risk. Full text: https://dedupedai.com/disclaimer/ URL: https://dedupedai.com/use-cases/hoylon-peerbridge-mcp/keep-approved-project-decisions-in-memory-ledger/