Protocol
Give a long AI task a durable feature ledger
A long prompt says what you wanted once. A ledger says what is still true now.
When it fits
- A coding or implementation task will span multiple sessions or agents.
When to avoid it
- Do not split tightly coupled work into fake independence; the ledger should reflect real acceptance boundaries.
Why it matters
Translate the target into end-to-end items with stable IDs, acceptance checks and status. Keep the ledger outside the chat so every resumed worker can see what is done, what remains and what is blocked. Mark an item complete only after its check passes; do not use the model's narrative of progress as the status source.
Steps
- A fresh agent can choose the next useful item without reconstructing the whole plan from conversation history.
An example
A site migration ledger lists routing, redirects, analytics and localization as separate outcomes with tests rather than asking an agent to 'finish the migration.'
Check your result
A fresh agent can choose the next useful item without reconstructing the whole plan from conversation history.
Keep this limit in mind
- Do not split tightly coupled work into fake independence; the ledger should reflect real acceptance boundaries.
Connected ideas
Useful withAsk a coding agent for one coherent increment at a time
Evidence and sources
Anthropic's long-running harness uses a structured feature list so each new session can see the remaining end-to-end work and avoid declaring the whole project complete prematurely.
A feature ledger is most useful when work can be decomposed into independently checkable increments.
Effective harnesses for long-running agents · Feature list and agent failure-mode table