Operating manual
Tag: ai-agents
A focused reading path for ai-agents: related field notes, evidence trails, and operating questions from the archive.
- parallel-tool-calls-need-a-resource-budget.md
Parallel Tool Calls Need a Resource Budget
A dependency graph shows which tool calls may overlap. A resource budget decides which can safely start with finite memory, browser slots, and API quota.
open artifact → - benchmark-agent-decisions-not-just-final-scores.md
Benchmark Agent Decisions, Not Just Final Scores
A final score hides whether an agent used evidence well. Evaluate the decision trace: context, intervention, consequence, validity, and budget.
open artifact → - adaptive-reasoning-choose-compute-not-defaults.md
Adaptive reasoning: choose compute, not defaults
Reasoning compute is a routing decision, not a default. Pick per-problem effort and prove the choice with token, latency, and accuracy gates.
open artifact → - the-missing-check-in-agent-handoffs.md
The Missing Check in Agent Handoffs
A practical two-sided contract for checking whether one agent's completed work leaves state the next agent can actually use.
open artifact → - a-world-model-is-an-executable-hypothesis.md
A World Model Is an Executable Hypothesis
A test-time world model should earn authority through replay, bounded action, and immediate revocation when reality produces a counterexample.
open artifact → - the-harness-is-becoming-part-of-agent-training.md
The Harness Is Becoming Part of Agent Training
An agent harness shapes what a model can see, do, retain, and optimize for. Use this six-question review to test training-to-production parity.
open artifact →