Business report org wide

Org-Wide Report (tier 1)

Project a measured per-instance saving across the organization and close with a summary the whole company can read. Tier 1: team-level aggregates only, no individual names anywhere. Value is engineer-hours and capacity; never money.

Read references/interim-format.md first — it defines run location and the completeness gate. Defer these rendering references until the gate passes; a refused render never loads them:

• references/provenance.md — tag rendering, anti-inflation rules, the no-currency contract.
• references/measurement.md — the projection chain V * F * t * s * a * r, the population segments, and the rule refusing 1.0 defaults.
• references/audiences.md — the tier-1 contract.

User arguments: $ARGUMENTS

Context

Recent runs — run this command and read the output:

sh -c 'ls -dt "$HOME"/Documents/*/business/ /mnt/c/Users/*/Documents/*/business/ 2>/dev/null' | head -5 | grep . || echo "(no runs found)"

Procedure

1. Locate the run

Use the $ARGUMENTS path if given, else the newest run per interim-format.md. Confirm the choice with the user.

2. Completeness gate

Run the completeness gate from interim-format.md. Refuse to render from an incomplete package; list what is missing and stop.

3. Secure the multipliers

Read the deferred references above. Then read adoption a and realization r from assumptions.yaml. If either is absent, ask the user via ask-user-choice — explain what each means per measurement.md — and append the answer to the register as an ESTIMATED entry with rationale and owner. Both are strictly below 1.0. Refuse to proceed with a default of 1.0 for either.

4. Render the report

Write <run>/reports/org-wide.md with this fixed structure:

1. Bottom-up build — the full chain Annual_hours_saved = V * F * t * s * a * r, every factor shown with its own tag; the population segmented addressable ⊃ served ⊃ realized, all three reported.
2. Scenarios — conservative, base, upside, with the differing assumptions listed per scenario.
3. Sensitivity — one-way ranking: which single input moves the outcome most, in order.
4. What has to be true — break-even framing: the values the inputs must reach for the projection to hold.
5. Plain-language summary — for the whole company: no methodology jargon, no formula notation; figures still tagged and true.

Team-level aggregates only. If any measurement is per-person, roll it up before it appears here.

Output

Open with a one-line hero (✓ Org-wide projection: <conservative figure with tag>, ✓ Evidence report: <supported claim, unknowns named>, or ⚠ Refused: <reason>), then exactly these sections:

1. ## Run — the run chosen and its pinned window.
2. ## Gate — the completeness result and the a/r values used, with their register ids.
3. ## Report — where it was written, the conservative projection, and the most sensitive input.

End with an ask-user-choice panel: write an internal case study next, revisit an assumption, or stop. Skip the panel in plan mode or when running non-interactively.

Portability notes

• ask-user-choice — present the listed options and wait for the user to pick one. Hosts with a structured multiple-choice tool (Claude Code's AskUserQuestion) should use it; otherwise print a numbered list and wait for a numbered reply. Never proceed on an assumed answer.
• $ARGUMENTS — the text the user passed when invoking this skill. If your host does not substitute it, read it as the user's request in the current turn, and ask when there is none.
• Bundled files — every relative path in this skill points at a file shipped inside this skill directory. Read them from here, not from the host's plugin tree.