Rebuild Dossier MCP server
{
"mcpServers": {
"parkerfawcett-rebuild-dossier": {
"command": "npx",
"args": [
"-y",
"rebuild-dossier@latest"
],
"env": {
"GROQ_API_KEY": "<groq_api_key>",
"MCP_AUTH_TOKEN": "<mcp_auth_token>",
"REBUILD_DOSSIER_TOKEN": "<rebuild_dossier_token>"
}
}
}
}{
"mcpServers": {
"parkerfawcett-rebuild-dossier": {
"command": "npx",
"args": [
"-y",
"rebuild-dossier@latest"
],
"env": {
"GROQ_API_KEY": "<groq_api_key>",
"MCP_AUTH_TOKEN": "<mcp_auth_token>",
"REBUILD_DOSSIER_TOKEN": "<rebuild_dossier_token>"
}
}
}
}{
"mcpServers": {
"parkerfawcett-rebuild-dossier": {
"command": "npx",
"args": [
"-y",
"rebuild-dossier@latest"
],
"env": {
"GROQ_API_KEY": "<groq_api_key>",
"MCP_AUTH_TOKEN": "<mcp_auth_token>",
"REBUILD_DOSSIER_TOKEN": "<rebuild_dossier_token>"
}
}
}
}{
"servers": {
"parkerfawcett-rebuild-dossier": {
"command": "npx",
"args": [
"-y",
"rebuild-dossier@latest"
],
"env": {
"GROQ_API_KEY": "${input:groq_api_key}",
"MCP_AUTH_TOKEN": "${input:mcp_auth_token}",
"REBUILD_DOSSIER_TOKEN": "${input:rebuild_dossier_token}"
}
}
},
"inputs": [
{
"id": "groq_api_key",
"type": "promptString",
"description": "Groq api key",
"password": true
},
{
"id": "mcp_auth_token",
"type": "promptString",
"description": "Mcp auth token",
"password": true
},
{
"id": "rebuild_dossier_token",
"type": "promptString",
"description": "Rebuild dossier token",
"password": true
}
]
}{
"mcpServers": {
"parkerfawcett-rebuild-dossier": {
"command": "npx",
"args": [
"-y",
"rebuild-dossier@latest"
],
"env": {
"GROQ_API_KEY": "<groq_api_key>",
"MCP_AUTH_TOKEN": "<mcp_auth_token>",
"REBUILD_DOSSIER_TOKEN": "<rebuild_dossier_token>"
}
}
}
}file .mcp.json (in your project root)file claude_desktop_config.jsonfile ~/.cursor/mcp.jsonfile .vscode/mcp.jsonfile ~/.codeium/windsurf/mcp_config.jsoninstall npx -y rebuild-dossierversion 0.2.6
What it does
An MCP server that analyzes an existing app and generates a locked rebuild specification—including interface contracts, configuration, and a mutation-tested test suite—that a coding agent can use to rebuild the app cleanly. Designed for developers who need to hand off app rebuilds to AI agents with verifiable specifications rather than guessing.
- Analyze an existing codebase to extract its routes, dependencies, and build configuration, then surface structural ambiguities that need resolution.
- Crawl a running app to document its actual behavior, then compare it against the inferred specification to identify discrepancies.
- Generate a locked rebuild spec with contracts and mutation-tested tests that enforce correct behavior when a coding agent rebuilds the app from scratch.
Tools this server exposes
| Tool | Does |
|---|---|
| ingest_repo | Static analysis of a repo: extracts routes, package.json, build config via AST, existing tests, and structural-smell detectors to surface ambiguities. |
| crawl_site | Headless Playwright crawl of reachable routes with progress notifications. |
| flag_known_bug | Store a free-text description of a known bug that overrides auto-resolution for matching ambiguities. |
| get_case_queue | Retrieve the queue of unresolved ambiguities that require human decision. |
| resolve_case | Record a human decision on an ambiguity by ID. |
| generate_spec | Generate CLAUDE.md, .claude/ config, spec/ contracts, and a mutation-tested test suite to a clean output directory; only callable once the case queue is empty. |
Before you enable it
An MCP server runs with whatever access you give it. Start with a development account rather than production, and check which tools the server exposes before you let an agent call them unattended.