Virtual Grounding Documents

BotDigit MCP Resources

While MCP Tools allow agents to take actions, MCP Resources allow agents to read structured context. BotDigit exposes custom botdigit:// virtual documents that keep AI models precisely aligned with real project realities.

Why MCP Resources Eliminate Prompt Hallucinations

When developers chat with AI models about client projects, they usually copy and paste lengthy requirements, chat transcripts, or outdated documentation into the prompt. This wastes expensive token context and risks catastrophic misunderstandings.

With BotDigit MCP Resources, the client IDE queries live virtual endpoints on demand. When an engineer in Cursor types @botdigit://workspaces/{id}/brief, the AI immediately accesses the authoritative project requirements, acceptance criteria, and architectural constraints directly from the source of truth.

Supported URI Schemes

botdigit://workspaces/{workspace_id}/brief

Workspace Project Brief

MIME: text/markdown

The authoritative client project specification, target architecture, user personas, and technical constraints.

Example Payload:

# Project Brief: Real-Time Payment Settlement Engine
## Objectives
- Build an idempotent payment reconciliation worker in Rust.
- Connect to Redis cluster at 6379 for fast de-duplication.
## Acceptance Constraints
- Zero duplicate payout dispatches.
- P99 latency under 25ms.

botdigit://workspaces/{workspace_id}/tasks

Active Kanban Board State

MIME: application/json

Current snapshot of all pending, in-progress, and completed tasks within the collaborative workspace.

Example Payload:

{
  "workspace_id": "ws_9921_alpha",
  "columns": ["todo", "in_progress", "review", "done"],
  "tasks": [
    { "id": "task_1", "title": "Implement webhook HMAC verification", "status": "done" },
    { "id": "task_2", "title": "Configure Stripe test webhooks", "status": "in_progress" }
  ]
}

botdigit://contracts/{contract_id}/milestones

Contract Milestones & Escrow Allocation

MIME: application/json

Contractual milestone definitions, dollar values held in escrow, expected deliverables, and deadline dates.

Example Payload:

[
  {
    "id": "ms_01",
    "title": "Phase 1: Architecture RFC & Database Migrations",
    "amount_usd": 2500,
    "status": "completed",
    "escrow_funded": true
  },
  {
    "id": "ms_02",
    "title": "Phase 2: Payment Gateway Integration & Tests",
    "amount_usd": 4000,
    "status": "in_progress",
    "escrow_funded": true
  }
]

botdigit://evidence/{milestone_id}/ledger

Cryptographic Milestone Evidence Ledger

MIME: application/json

Audit trail of linked Git commits, PR merge events, and verification logs proving acceptance criteria satisfaction.

Example Payload:

{
  "milestone_id": "ms_01",
  "audit_hash": "e3b0c44298fc1c149afbf4c8996fb92427ae41e4649b934ca495991b7852b855",
  "events": [
    { "event": "git_commit_linked", "sha": "4a18f92", "author": "dev_alex" },
    { "event": "test_suite_passed", "tests_run": 84, "failures": 0 }
  ]
}

botdigit://talent/{username}/passport

Verified Talent Passport

MIME: application/json

Publicly verifiable freelancer profile with cryptographically signed skills, completion history, and rating index.

Example Payload:

{
  "username": "rust_architect",
  "passport_id": "tp_88190_botdigit",
  "skills": ["Rust", "Actix-Web", "PostgreSQL", "Kafka"],
  "projects_completed": 38,
  "on_time_delivery_rate": 0.99
}

Real-Time Resource Subscriptions

BotDigit MCP supports the protocol's resources/subscribe notification primitive. When a client stakeholder updates milestone acceptance criteria on the web, connected AI agents receive instantaneous push notifications over their active SSE stream, ensuring the agent never builds against stale requirements.

How Autonomous Agents Work

Explore how multi-agent swarms leverage BotDigit MCP.

Autonomous Agents Guide

Related MCP Architectural Guides

17 MCP Tools Catalog

Executable actions for agents to interact with workspaces and tasks.

Autonomous AI Agents

How autonomous and supervised AI agents operate within BotDigit.

Security & Scopes

Resource-level authorization and workspace privacy filters.