---
name: Executive Briefing Optimization
description: From being overwhelmed by information to efficiently understanding key points and actions.
velocity:
  skillId: skl_marketing_handout
  activityCode: ACT-000103
  activityName: Executive Briefing Optimization
  version: 1
---

# Executive Briefing Optimization
Turn scattered streams of inbound information into a tight, decision-ready executive briefing that surfaces only what You needs to know and do.

## When to use this skill
Invoke this skill when You (CEO) feels buried in updates — project threads, status reports, team discussions, deadlines — and needs a clean read on what matters now. Use it for a daily/weekly morning brief, before a leadership meeting, or any time the request is "catch me up," "what needs my attention," or "summarize where things stand." It is built to run on a recurring schedule: You can ask Claude to schedule it (e.g. every weekday at 7am).

## Required connections
- Cowork

## Instructions

1. **Establish scope.** Determine the time window (default: since the last briefing, or last 24 hours for a daily brief) and the set of projects/teams to cover. If You named specific initiatives, restrict to those; otherwise cover all active projects.

2. **Pull project state from Cowork.** Use the Cowork API to retrieve current project tasks, assignments, deadlines, and progress tracking across the in-scope projects. Capture: tasks that changed status, items overdue or due within the window, blocked items, and owner per item.

3. **Pull communication signal from Cowork.** Use the Cowork API to generate summaries of recent team discussions and upcoming meetings within the time window. Extract decisions made, open questions raised, and any item where a team explicitly flagged it as needing CEO input or approval.

4. **Triage ruthlessly.** Classify every item into exactly one bucket:
   - **Decide** — needs You's explicit decision or approval.
   - **Aware** — material change You should know but need not act on.
   - **Risk** — slipping deadline, blocker, or escalation trending wrong.
   - **Noise** — drop it. Do not surface routine, on-track, or low-signal items.
   Keep only Decide / Aware / Risk in the brief. The whole point is reducing volume — when in doubt, demote toward Noise.

5. **Rank by leverage, not by recency.** Order surviving items by impact on revenue, customers, team, or strategic commitments. A CEO brief leads with the highest-stakes item, not the newest one.

6. **Draft the briefing** in this fixed shape so You can scan it in under two minutes:
   - **TL;DR** — 3 bullets max, the single most important thing in each of Decide / Risk / Aware.
   - **Decisions needed** — each as: *the ask* → *the recommendation* → *the deadline*. One line each. If you can recommend a default, do.
   - **Risks & blockers** — what's slipping, why, who owns it, and the suggested intervention.
   - **For awareness** — terse one-liners, no elaboration.
   - **Nothing-burger note** — one line confirming what was reviewed and found on-track, so You trusts the omissions (e.g. "All other 11 active projects on track; no action needed").

7. **Make every item actionable.** For Decide and Risk items, propose the next action and, where possible, draft it — e.g. a suggested reply, a reassignment, or a deadline change. Do NOT execute changes in Cowork automatically; present them for approval unless You has pre-authorized a specific action.

8. **Deliver and offer follow-through.** Present the briefing, then ask whether You wants any approved actions enacted in Cowork (task reassignment, deadline update, automated reminders to owners) using the Cowork API workflow automation.

9. **Offer to schedule.** If this was a one-off run, ask whether You wants it on a recurring cadence and at what time. Confirm the schedule, scope, and delivery format before setting it.

## Constraints
- Hard ceiling: TL;DR ≤ 3 bullets, Decisions ≤ 5 items, total brief readable in ~2 minutes. If more than 5 decisions exist, surface the top 5 and note the count of deferred items.
- Never pad. An empty section is good news — render it as "None this cycle."
- Attribute every Decide/Risk item to a source (project or thread) so You can drill in.
- Respect Cowork rate limits — batch retrievals rather than per-item calls.

## Example

**User:** "Give me my morning brief."

**Response shape:**

> **TL;DR**
> - 🟠 **Decide:** Acme renewal terms need sign-off by EOD.
> - 🔴 **Risk:** Platform migration slipped 1 week — engineering blocked on vendor.
> - 🔵 **Aware:** Q3 hiring plan finalized by People team.
>
> **Decisions needed (2)**
> 1. *Approve Acme 3-yr renewal at 8% uplift?* → Recommend **yes**, in line with target margin → due today 5pm.
> 2. *Greenlight $40k spend on migration vendor?* → Recommend **yes**, unblocks 4-person team → due tomorrow.
>
> **Risks & blockers (1)**
> - Platform migration: vendor SLA dependency stalling rollout. Owner: Priya. Suggested move: approve item #2 above to clear it.
>
> **For awareness (2)**
> - Q3 hiring plan locked — 6 reqs opening Monday.
> - Customer NPS up 4 pts week-over-week.
>
> _Reviewed all 9 active projects; 6 on track, no action needed._
>
> Want me to enact the approved decisions in Cowork, or schedule this brief for 7am every weekday?

## Working folder

Save what you produce into the learner's **Velocity Work** folder for this activity — Velocity sets one up named "Executive Briefing Optimization" inside their Velocity Work folder (by default `~/Documents/Velocity Work/Executive Briefing Optimization`). Read their own input files from the folder they connected to this chat. If that folder looks empty, or you're unsure which one to use, ask the learner to confirm or connect the Velocity Work folder for this activity before you write — so their work lands where Velocity (and any runs they schedule later) can find it.

## Recording the invocation

When you start this skill, call the **velocity-receipts** MCP server's `record_invocation` tool with this skill's `skillId` (`skl_marketing_handout`, from the frontmatter above — not the slug) and your starting context. When you finish, call `complete_invocation` with the invocation id, a one-line summary of what you accomplished, and a **dataContext**: pass `metadata: { dataContext: "real" }` when you worked on the learner's OWN data (their Velocity Work folder for this activity) — that records real Value Impact — or `metadata: { dataContext: "practice" }` when you worked on the provided practice data, which is logged as practice and kept out of their Value Impact total. If you're unsure, ask the learner which they used. The MCP server writes a SkillReceipt that compounds into your Value Impact total — no manual recording needed.

## Scheduling this skill

This skill is built to run on a recurring schedule. To set that up, create a **Claude scheduled task** that invokes this skill on the cadence the user wants (for example, every weekday at 8am). The scheduled task should run the skill exactly as you would manually: call `record_invocation` at the start and `complete_invocation` at the end. When the run is scheduled (not a manual invocation), pass `metadata: { scheduled: true }` to `complete_invocation` so the run is credited as a scheduled one. From then on the skill keeps recording Value Impact on its own — the user doesn't need to be present.

---
*Velocity skill · ACT-000103 · v1*