---
name: analyst-handoff
description: Capture analytics knowledge from an analyst who is changing roles, leaving, or transferring domain ownership, producing governed ACF context and eval seeds through context-interview. Use only for an intentional knowledge-transfer or analyst-transition request.
---

Approval contract: new and changed context stays on a named working branch and
has no authored approval status. Human-reviewed protected merge grants definition
approval; consult the live authority through the runtime for the exact revision
and freshness. Missing/expired authority evidence prevents governance certification;
it does not by itself prohibit an explicitly approved, unverified read-only analysis.
Known revoked definitions or materially stale dependencies still require resolution.
Explicit branch-test mode is proposed context only. Query/notebook result
verification is separate from definition approval; retain source and review
evidence. See the context-interview SPEC approval-through-merge contract.


# Analyst Handoff

Run a time-bounded, transition-focused form of `context-interview`. The departing
analyst's knowledge must become reviewable context and ground-truth examples, not
a transcript that another analyst has to rediscover later.

## Dependency and ownership

Use the sibling `context-interview` skill for repository discovery/scaffolding,
the confirmation ladder, ACF authoring, eval-seed harvesting, live verification,
and validation. If that skill is unavailable, stop and ask the user to install the
complete Nodal Analytics package; do not invent a second ACF workflow.

The human remains the owner of every confirmed definition. Warehouse and dashboard
access remain read-only. Never request or handle credentials.

## Handoff overlay

1. Establish the transition date, available session time, domains being handed
   off, current owner, and intended successor or escalation role. Do not infer a
   company or person's identity from schema names.
2. Locate or create the context repo through `context-interview`, then read its
   current domain roster and open drafts. Prefer finishing high-risk existing
   drafts before opening low-value new domains.
3. For each domain, apply the prompts in `references/handoff-checklist.md` through
   the normal selection/confirmation ladder. Work in batches of one to three and
   show progress.
4. Every confirmed disambiguation emits an eval seed. Record rejected credible
   readings in intent and shape exclusions. Use a negative-kind seed only for a
   correctly declined/escalated question, with its declared reason. Unconfirmed recollections remain drafts with a
   named follow-up owner or role.
5. Run the normal per-domain live verification where time and tools allow. A
   missing connection degrades to semantic capture; it never turns recollection
   into verified truth.
6. Validate the context repo and finish with a handoff coverage report.

Carry explicit open decisions, method decision rationales, family/alias choices,
stale seeds and pending Stage 6 observations into the successor's agenda. Notebook
knowledge can become a statically imported draft method, followed by human review
and independent Stage 5 verification when authorized and available. Do not treat
an import, a simulation or a departing analyst's recollection as an execution.
Keep raw observations and execution evidence local; confirmed metric/method/seed
records preserve the actual approver and date.

## Coverage report

Report by domain:

- confirmed definitions and eval seeds added;
- remaining drafts and why they are unresolved;
- dashboards/reports and operational deadlines;
- known wrong approaches and data-quality traps;
- successor or escalation owner;
- verification completed, unavailable, or still required.

Do not claim the handoff is complete while a load-bearing domain has no successor,
contains unresolved contested definitions, or lacks a clearly named follow-up.