News & guides
Orchestrator-Worker Patterns for Agents That Hold Up
The orchestrator-worker shape is the one multi-agent design that reliably pays for itself. Here is how to build the lead, the brief and the synthesis step.
· 9 min read
Parallel Tool Calls: When Concurrency Helps an Agent
Running tool calls concurrently cuts wall-clock time but not tokens, and it is only safe for some operations. How to decide what to parallelise.
· 8 min readPartial Failure Recovery: When Half the Agents Succeed
Six workers, four returned, two timed out. Whether to answer, retry or abort is a design decision — here is how to make it before the incident, not during.
· 9 min readPrompt Regression Testing: A CI Gate for Fuzzy Output
Prompt edits regress silently and every score is noisy. How to build a golden set, pick a scorer, handle variance and gate merges without a permanently red build.
· 9 min readRate Limiting Agent Fleets Without Killing Throughput
One API key, many agents, one shared quota. How to build a fleet-wide limiter that respects token budgets, protects interactive work and survives bursts.
· 9 min read
Replaying Agent Traces: Debugging a Run You Cannot Reproduce
Agent runs rarely reproduce, so the stored trace is the only evidence. What to capture, how to replay it, and where replay stops being faithful.
· 9 min read
Rollback Strategies for Agents That Changed Things
An agent stopped halfway through eleven changes. Checkpoints, compensating actions, staging areas and the undo plan you should require before it acts.
· 9 min readScratchpad Patterns: Giving an Agent Somewhere to Think
A durable scratchpad survives compaction, crashes and handoffs. What belongs on it, what does not, and how to stop it becoming a second transcript.
· 8 min readShared Memory Between Agents: Designs That Actually Hold
Agents that share a scratchpad drift, overwrite and confuse each other. Here are four shared-memory designs, what each one breaks, and how to pick.
· 9 min readSpeculative Agent Execution: Running Ahead of the Decision
Agents spend most of their wall-clock time waiting. Speculative execution starts likely next steps early and discards the wrong guesses. Here is when it pays.
· 9 min read
Summarisation in Agent Loops: Compressing Without Forgetting
Long sessions have to shed context. What to summarise, what must survive untouched, when to trigger it, and how to tell a good summary from a lossy one.
· 8 min read
The MCP Security Model: Trust Boundaries You Actually Have
MCP standardises connection, not trust. Here are the real boundaries in an MCP deployment and the attacks that cross them, with practical mitigations.
· 9 min readReady to try it?
Every model behind one key. Start free and upgrade when it sticks.