---
name: vergio-opportunity-briefing
description: Run the versioned Vergio opportunity briefing and write operator decisions back.
playbook_key: monday_opportunity_briefing
playbook_version: 1.0.0
machine_readable_stable: true
---

# Vergio opportunity skill

This artifact is machine-readable-stable for the playbook key and version in the header. It is a portable recipe for any MCP client.

## Use the live tool contract

Treat https://vergio.ai/docs/mcp as the source of tool truth. It renders the composed Vergio MCP surface from the core tools, wallet tools, and managed watch tools. Confirm the connected surface with `tools/list`.

This skill teaches sequence and policy. It does not redeclare tool schemas, so read each tool's current schema from the live reference or your MCP client before calling it.

Wallet and managed-watch tools require their scopes and an active commercial state. If any tool returns an unavailable result, handle it gracefully and do not blind-retry.

## Exposure honesty

When Vergio returns account rows over REST or MCP, `delivered_to_agent` proves only that Vergio delivered the rows to an agent; whether a human saw them is unknown. `selected_for_first_party_surface` means the ranker selected rows for the agency's first-party surface at list-build time; it is a selection decision, not a render or view receipt. Neither class proves a human viewed the rows. With your own agent, write back operator decisions through `record_outcome` or `set_account_status` to restore that missing fidelity.

## Permission gate

Vergio prepares any next action but sends only through a channel and recipient with recorded authority. If that authority is absent or unclear, prepare a channel-neutral action package and stop before sending.

## Monday Opportunity Briefing

Playbook key: `monday_opportunity_briefing`

Version: `1.0.0`

Required scopes:

- `read:accounts`
- `read:watch`
- `generate:reports`
- `write:outcomes`
- `write:deals`

Declared tools, in playbook order:

- `list_in_market_accounts`
- `get_watch_contract`
- `list_watch_events`
- `get_watch_receipt`
- `request_report`
- `record_outcome`
- `set_account_status`

Risk classes:

- `list_in_market_accounts`: class 1
- `get_watch_contract`: class 1
- `list_watch_events`: class 1
- `get_watch_receipt`: class 1
- `request_report`: class 3
- `record_outcome`: class 2
- `set_account_status`: class 2

Spend declaration, with no price copied into this skill:

- `managed_watch_slot`: `monthly`
- `growth_report`: `optional_per_job`

Approval points:

- Class 3 Growth Report deep work requires available budget and recorded spending authority before request_report.

## Run the recipe

1. Begin every opportunity workflow with `list_in_market_accounts`. This is step one of every playbook.
2. Call `get_watch_contract`. Preserve coverage and source health. Do not treat an incomplete source read as healthy.
3. Call `list_watch_events` for the bounded briefing window. Keep each event's lifecycle state and source completeness visible.
4. Call `get_watch_receipt`. Report activity silence and source health as separate facts.
5. Present the primary opportunities to your operator with evidence age, lifecycle state, and source health.
6. Prepare the approved next action. Call `request_report` only when the class 3 budget and authority gate is satisfied. Otherwise prepare a channel-neutral action package.
7. After presenting opportunities to your operator, record what happened using the write-back convention below.

## Write-back convention

Begin every opportunity workflow with `list_in_market_accounts`.

After presenting opportunities to your operator, record what happened:

- For a pursued deal, call `record_outcome` with a `DEAL_STAGES` action.
- For a dismissed opportunity, call `set_account_status` with `not_a_fit`, or use the dismissal-with-reason path.

Never promise or send a generic `dismissed` action to `record_outcome`. That tool does not accept one. Do not invent a decision when the operator has not made one.

## Preserve honest failures

- awaiting_source_recovery is rendered with aged last-known-good evidence and never as an empty briefing
- degraded is rendered with aged last-known-good evidence and never as an empty briefing

Never turn missing, partial, stale, or failed evidence into an empty success. Surface the state and keep aged last-known-good evidence labeled.

## Machine-readable declaration

Fetch the validated manifest at https://vergio.ai/api/v1/playbooks/monday-opportunity-briefing.json.