Skip to main content
This section is the internal operating handbook for AI agents and developers.

Goals

  • fast onboarding for new agents
  • predictable change safety gates
  • consistent documentation updates as the code evolves
  • explicit ownership for manager/support agent workflows

Internal Documentation Structure

  • internal/overview: start here
  • internal/manager-support-model: role contract and merge gates
  • internal/change-protocol: mandatory preflight, during-change, and postflight checklist
  • internal/docs-maintenance: how to keep docs current without drift
  • internal/skill-blueprint: reusable skill template for DealDash-specific agent behavior
  • internal/dealdash-agent-bridge-plugin: provider-neutral MCP bridge and approval rules
  • internal/openclaw-channel-node-policy: locked communication and execution policy
  • internal/openclaw-infrastructure-optimization: production topology and hardening plan
  • internal/openclaw-learning-loop: controlled autonomy learning loop

Canonical Sources Inside DealDash Repo

  • /AGENTS.md
  • /docs/agents/README.md
  • /docs/agents/OPENCLAW_MANAGER_PLAYBOOK.md
  • /docs/agents/SUPPORT_AGENT_PLAYBOOK.md
  • /docs/agents/CHANGE_SAFETY_PROTOCOL.md
  • /docs/agents/TASK_ROUTING_MATRIX.md
  • /docs/agents/DEALDASH_AGENT_SKILL.md
  • /docs/agents/DEALDASH_AGENT_BRIDGE_PLUGIN.md
  • /docs/agents/OPENCLAW_CHANNEL_NODE_POLICY.md
  • /docs/operations/OPENCLAW_INFRA_OPTIMIZATION_PLAN.md
  • /docs/agents/OPENCLAW_LEARNING_LOOP.md

Done Definition For Every Engineering Task

  1. behavior change implemented safely
  2. relevant checks pass
  3. docs are updated in the same change
  4. residual risk is written down