/rollup

name: rollup
description: Generate the federation roll-up catalogue for a multi-repo product — aggregate every member repo's ADR metadata into one derived, product-wide view, run from the home repo. Use when the user says "roll up the federation", "generate the product-wide ADR view", "aggregate ADRs across the repos", "refresh the roll-up", or invokes /rollup. NOT for regenerating a single repo's own INDEX (that happens in place during authoring/ship) and NOT for linting consistency (use the audit skill).

rollup

Aggregate the ADR catalogues of every repo in a multi-repo product into one **derived, read-only** product-wide view. The roll-up is regenerated from source, never hand-edited — treat it exactly like a repo's own `INDEX.md`.

Step 0 — Preconditions

1. This skill runs in the **index-holding repo** of a federation — the one that carries `federation-index.md` (its `Role` is `central` for topology A, `coordinator` for B, or `home` for C). Confirm `federation-index.md` and a `federation.md` whose `Role` is `central`, `home`, or `coordinator` exist. If they do not, stop: either this repo is standalone (nothing to roll up) or it is a plain member — point the user at the index-holding repo. 2. Read `federation.md` to learn the **identity scheme** (default repo-prefixed slug `<repo-id>/NNNN-slug`); roll-up rows use it.

Step 1 — Enumerate members

Read `federation-index.md`. For each row, take the `Repo id` and the `Pointer` (the path/URL of that member's checkout). The member index is the **only** source of membership — do not auto-discover repos.

Step 2 — Collect each member's catalogue

For every member whose checkout is **locally available** at its pointer:

• Read that member's `INDEX.md` (its authoritative local catalogue).
• For each ADR row, capture the number, title, status, date, and

dependencies, and attach the **owning repo id** and the **federation identity** (the scheme from Step 0 applied to the local number).

A member's own `INDEX.md` stays authoritative for that member; this skill only reads it.

For the **aggregate status** of a product-wide decision (one with owning per-repo plan items across several members), also scan each member's `plan/todo/` (pending) and `plan/done/` (shipped) for items naming that decision's federation identity — that per-repo state feeds the aggregate column in Step 4.

Step 3 — Handle unreachable members

A member named in the index whose checkout is **not locally available** is **not** dropped and **not** a failure. Record it in a clearly separated "Not aggregated this run" list with its repo id and pointer, so the gap is visible rather than silent.

Step 4 — Write the roll-up

Write the aggregate to **`ROLLUP.md`** at the configured artefact root (the same root as `INDEX.md`), so every re-run overwrites the same file:

• A header stating it is **generated — do not hand-edit**, with the run

date and the set of members aggregated.

• One table across the whole product: federation identity, title, status,

owning repo, date. Group or sort by owning repo for readability.

• For product-wide decisions, an **aggregate status** column **derived**

from per-repo plan-item state: `Implemented` only when every owning per-repo plan item is in a member's `plan/done/`, otherwise `N of M repos` shipped. This column is derived — never written back into any ADR.

• The "Not aggregated this run" list from Step 3.

Do not alter any member's `INDEX.md` and do not write into any other repo.

Step 5 — Report

Tell the user how many members were aggregated, how many were skipped as unreachable (and which), and the total ADR count in the roll-up. Remind them the file is derived: re-run this skill to refresh it rather than editing it by hand.

Name catalogue coverage and unavailable members; partial coverage is not verification.

<!-- docflow:closing-report -->

Closing report

End every run, including blocked, failed and stopped runs, with a section headed exactly **Status at a glance**, containing these three labels:

• **This run:** only actions actually attempted and their outcomes; quote each verify gate's exact output and exit code, including timeouts or interruptions.
• **Overall:** implemented, partially verified, verified, blocked, failed or unknown. A passing sub-step is not an overall pass; incomplete or missing evidence never becomes success.
• **Yet to do:** every remaining action, unresolved finding, verification, cleanup or required input. Write None only when the whole task is verifiably complete; never omit work because a budget ended.

Routine progress messages need no block. Keep final results brief and distinguish work prepared on a PR from work confirmed shipped. <!-- /docflow:closing-report -->