Validate agent handoff packets and resume readiness using schema, freshness, and replay checks. Use when tasks pause/resume across sessions, agents, or humans — including when a user wants to continue where they left off, hand off to another agent, resume a previous task, or pick up an interrupted workflow. Includes explicit untrusted-content/prompt-injection guardrails for third-party inputs.
100
100%
Does it follow best practices?
Impact
100%
1.31xAverage score across 3 eval scenarios
Advisory
Suggest reviewing before use
A data-engineering team was mid-way through setting up an automated ETL pipeline when the original developer had to take unexpected leave. The project was put on hold for several weeks, and the team has now reassigned it to a new contractor. Before the contractor begins work, the project manager wants a proper integrity check on the handoff packet that was left behind, so they know exactly what state the work is in and what steps are needed before it's safe to continue.
The handoff packet was drafted hurriedly and the project manager suspects it may have some issues — but needs a clear, formal report rather than informal opinions. Produce a structured validation report that the team can use to decide how to proceed.
Write your validation report to handoff-audit.md. The report must clearly indicate the status of each check performed, an overall health classification, the specific steps that would be needed to get the handoff into a resumable state, and guidance on whether the task owner needs to be involved before work restarts.
Also produce a remediation-plan.md that lists the concrete changes needed to fix the handoff packet before it can be safely used, including a suggested replacement for any problematic fields.
The following files are provided as inputs. Extract them before beginning.
=============== FILE: inputs/handoff.json =============== { "objective": "Build and deploy the customer-events ETL pipeline to production", "completed": [ "Defined source schema for customer_events table", "Wrote initial Spark job for event transformation" ], "unresolved": [ "Integration tests not written yet", "Production deployment config not finalised" ], "assumptions": [ "Cloud budget approved for staging environment" ], "next_action": "Write integration tests covering the three main event types before merging the Spark job PR", "risks": [ "Schema drift from upstream if source team makes changes before we deploy" ], "updated_at": "2026-01-05T09:00:00Z", "resume_token": "Bearer eyJhbGciOiJIUzI1NiIsInR5cCI6IkpXVCJ9.eyJzdWIiOiJldGwtcGlwZWxpbmUtam9iIiwiaWF0IjoxNzA0NDQ0MDAwfQ.abc123signature" }