Multi-agent coordination

Parallel and multi-agent coordination

• How to Run Many AI Coding Agents in Parallel Safely
• A Fixed Review Order for Ten Finished Agent Branches
• When Two Agents Disagree: Resolving Conflicting Outputs
• A Changed-Lines Registry for Multiple Agents
• A Lockfile Diff Budget for Multi-Agent Branches
• Costing a Three-Agent Loop: Tokens and Latency Per Conflict
• How to Split a Task Across Two Agents When One Pass Is Not Enough
• Splitting the Test Writer from the Implementer
• All of Multi-agent coordination

Task queues

Agent task queues and handoffs

• Who Owns This Branch? Tracking Agent Work Across a Queue
• A Handoff Protocol for Agent Tasks
• A Patch Queue for Agent Work: Splitting a Large Diff
• A Queue Record Schema for Unmatched Requests
• Claimed, Resolved, Requeued: Queue States and Timers
• Dead Letters for Agent Tasks: Where Failed Work Goes
• From Checks to a Queue: What an Agent Picks Up Next
• From Self-Review to a Queue: What an Agent Picks Up Next
• All of Task queues

Approval workflows

Human approval and escalation workflows

• Evidence Links That Survive a Rerun
• From Override to Rule Change: A Pull Request for the Prompt
• How Much Context a Batched Approval Item Needs
• Reading the Proposal: A Queue for Agent Writes Before Apply
• Routing Misses to a Human Queue: Fields and Timers
• The Approval Queue: Routing and Expiring Decisions
• A Ledger for Dependency Overrides and Their Approvals
• A Path-to-Reviewer Rules Table for Agent Branches
• All of Approval workflows

Agent stopping rules

Agent decision-making and stopping rules

• Diff Budgets: Setting Hard Limits on Agent Changes
• Prompt Patterns for Stopping an Agent Mid-Task
• When an Agent Should Stop and Ask a Person
• A Stop Rule That Fires Too Often: Measuring Over-Asking
• Encoding a Refusal: Declining a Task Not to Start
• Instrumenting Agent Halts: Metrics for Stop Rate
• Setting a Time Budget for Each Agent Command
• Splitting Decisions: What an Agent May Settle and What It May Not
• All of Agent stopping rules

Skill routing

Routing requests and skills to agents

• A Trigger Regression Set for Skills: Sampling and Pass Criteria
• Building a Tie-Breaker Agent: Rules and Risks
• Routing to Nothing: When the Right Answer Is No Skill
• When a Router Picks a Tool the Agent Cannot Call
• Writing a Skill Router: How Models Choose Between Skills
• A Canary Description for Skill Routing: Shadow Loads
• A Confidence Band That Survives a Model Swap
• A Confusion Table per Skill: Reading False Loads and Misses
• All of Skill routing