Three-component handoff spec: output format, explicit expectations, and return protocol — ambiguous handoffs create bottlenecks
For each handoff between agents or pipeline stages, specify three components—defined output format, explicit expectations, and return protocol—to prevent ambiguous handoffs from creating bottlenecks.
Why This Is a Rule
Handoffs — the points where one person, process, or agent passes work to another — are where most workflow failures occur. Not because the upstream work was bad or the downstream work was bad, but because the transition between them was ambiguous. The upstream agent delivered "something" and the downstream agent expected "something else," producing rework, delays, or misalignment.
Three components eliminate handoff ambiguity: Defined output format: what exactly does the upstream agent deliver? Not "a draft" but "a 500-word draft with section headers, in Google Docs, with comments enabled." The specificity prevents format mismatches. Explicit expectations: what quality standards must the output meet before handoff? Not "ready for review" but "passes spell check, all claims have citations, logical flow verified." This prevents premature handoffs that produce downstream rework. Return protocol: what happens if the downstream agent needs something different? Not "figure it out" but "flag specific issues in comments, return to upstream within 24 hours with the single highest-priority revision needed." This prevents handoffs from becoming black holes.
This extends Specify the information contract at every agent handoff — what exact output does the next step need from the current one? (information contracts) with the quality and return dimensions that contracts alone don't cover.
When This Fires
- When designing any collaborative workflow where work passes between people, teams, or stages
- When handoff points consistently produce delays, rework, or misunderstanding
- When delegation (Every delegation needs three written components: one accountable owner, authority constraints, and escalation triggers) produces outputs that don't match expectations
- When pipeline stages have unclear boundaries about what "done" means for handoff purposes
Common Failure Mode
Implicit handoffs: "Send it to Sarah when you're done." What format? What quality bar? What if Sarah needs changes? All three are implicit, producing: wrong format (Sarah expected a spreadsheet, got a document), inadequate quality (the work needs another editing pass before Sarah can use it), and no return protocol (Sarah sits on it for a week trying to make it work rather than returning it for revision).
The Protocol
(1) For each handoff point in a workflow, specify three components: Output format: exact deliverable description. What file type, structure, content elements, and naming convention? What does "done" look like in concrete terms? Expectations: quality criteria the output must meet before handoff. Checklist of minimum requirements. The upstream agent self-verifies against this list before handing off. Return protocol: if the downstream agent can't use the output as-is, how do they communicate what's needed? What's the expected turnaround for returns? What format do return requests take? (2) Document all three alongside the workflow (Map agent dependencies and arrange in topological order — no agent executes before its required inputs are available). (3) At each handoff, the upstream agent verifies against the expectations checklist before delivering. The downstream agent verifies receipt against the format spec. Mismatches are caught at the handoff rather than downstream.