---
name: skill-opportunity-eval
type: skill
description: Evaluate a new opportunity — a venture, a role, a bet, a partnership — against current strategy, goals and capacity, and produce a structured decision brief. Use this whenever Aled is weighing whether to take something on, asks "should I do X", or wants a clear read on a new bet before committing. Produces a brief covering what it is, fit against the active strategy and initiatives, cost in time, money and attention, risk, and a clear recommendation with its reasoning and the opposing case. Decision support only — never commits or starts work.
license: Proprietary — Aled Pritchard workspace use.
---

# opportunity-eval

Atlas's decision-support skill. When a new opportunity lands — a venture to start, a role to take, a bet to place, a partnership to enter — this turns it into a structured brief Aled can decide from: what it is, how it fits the strategy, what it costs, what it risks, and a clear recommendation that also argues the other side. The output is a brief, never an action. Runs on the Linear connector only. Follows `convention-linear` and `core-operating-model` throughout.

## Trigger

Aled describes an opportunity and asks for a read, or asks whether to take something on. Cross-domain — the opportunity may touch any venture (madebyneud, careerOS, A1, Mr Pritchard, the apps) or sit outside them entirely. Never touches the Pipeline team's own workflow; this is strategy support, not pipeline qualification.

## Behaviour

A run has three passes, in order.

### Pass 1 — Understand the opportunity and the context

Capture what the opportunity actually is, in Aled's terms — its shape, what it would require, and what it promises. Then read the current strategy context: the active outcome-initiatives and their KPIs (from `skill-strategy-cascade`), the current focus recommendation (the latest `skill-plan` output), and a realistic read of present capacity. The brief is only as good as the context it is weighed against.

### Pass 2 — Assess

Work through, plainly:

- **What it is** — a tight description, no inflation.
- **Fit** — how it sits against the active strategy and initiatives. Does it advance a stated goal, sit adjacent, or pull against them?
- **Cost** — time, money, and attention. Attention is the scarce one; name what it would displace.
- **Risk** — what could go wrong, what is uncertain, and what the downside looks like if it does not work.

### Pass 3 — Recommend, with the opposing case

Give a clear recommendation — take it, decline it, or a conditional ("take it if X") — with the reasoning that leads there. Then state the opposing case honestly: the strongest argument against the recommendation. A brief that only argues one side is not decision support; the opposing case is what makes the call trustworthy.

Post the brief for Aled, assigned to him. It is a brief to decide from — this skill never commits, books, or starts work off the back of it.

## Decision brief

Post to the relevant venture project, or the os.Claude project where the opportunity is cross-domain:

```
[opportunity-eval] [name] — [date]

What it is:
[tight description]

Fit against strategy:
[advances / adjacent / pulls against — which initiatives, why]

Cost (time / money / attention):
[what it takes, what it displaces]

Risk:
[what could go wrong, the downside]

Recommendation:
[take / decline / conditional] — [reasoning]

The opposing case:
[strongest argument against the recommendation]
```

## Guardrails

- Decision support only. Never commits, books, starts work, or creates initiatives off the back of the brief.
- Always argues the opposing case — never a one-sided brief.
- Names attention cost honestly, including what the opportunity would displace.
- Linear only — never writes code, opens a PR, or touches the delivery loop.
- Aled is the gate. Atlas proposes the read; Aled makes the call.

## Setup

- Run on demand when Aled is weighing an opportunity. Connector: Linear only.
- Not scheduled — opportunities arrive on their own clock.

Owned by Atlas (ATL-4S).

On finish, run skill-ops-retro capture on this run's friction: file a FRICTION note to the Pulse queue for the drain pass to assess.