Understanding Context Continuity
What re-explanation rate and receipt reuse tell you about whether your team is burning time rebuilding context that already exists.
The hidden cost of re-explanation
Every time a team member starts a new AI coding task and has to re-explain what was done in a related prior task, that is waste. Not the task itself, not the code, but the overhead of reconstructing context that was already captured and then discarded.
The Context Continuity report measures this overhead across the team. It tracks how often prior session receipts are reused as starting context for the next run, and how often context is rebuilt from scratch instead.
Receipts reused
When a task starts and Avorelo finds a relevant prior receipt, it uses it. The prior run's verified output, scope record, and findings become available to the new run without re-reading the same files or re-establishing the same context.
A high receipts reused count means your team's prior work is carrying forward. A low count can mean that related tasks are not being linked through receipts, that the prior runs are producing unverified exits that cannot be safely reused, or that team members are starting new sessions instead of continuing chains.
Re-explanation rate
This is the proportion of runs that started without finding a usable prior receipt, and had to establish context from scratch. It includes both cases where no prior receipt existed and cases where a prior receipt existed but was not clean enough to reuse.
A re-explanation rate above 40% suggests that the receipt chain is not holding together. Common causes: unverified exits from prior runs, tasks that are not structured as continuations of prior work, or chains that span too many days for the receipt to remain relevant.
The re-explanation rate is also a token cost signal. Each rebuilt context uses tokens to re-establish what a reused receipt would have provided for free. The context reuse savings in the metrics section estimates what this costs over the period.
Context reuse by source
The table in the report shows individual handoff events: which prior receipt was reused, which task benefited from it, and whether the reuse was automatic (receipt was clean and relevant) or manual (team member explicitly referenced prior work in the task description).
Rebuilt rows are the actionable ones. Each rebuilt row is a receipt that should have been reusable but was not. Click through to see why: unverified exit from the prior run, task chain gap, or a scope mismatch between the prior run and the new task.
What to improve
- Reduce unverified exits (see Workflow Health report) to increase the pool of reusable receipts
- Structure related tasks as explicit continuations when starting a new session on the same work area
- Review the rebuilt rows for patterns: if the same area of the codebase keeps requiring context rebuilds, that area may benefit from a persistent context note
Open Context Continuity report
View receipt reuse and re-explanation metrics for your team