---
name: loomi-project-discovery-brief
description: Build a compact Bloomreach Engagement project intelligence brief from local contexts, supplied projectContextJson, Loomi MCP discovery when needed, and optional Bloomreach CLI workspace cards when available. Use when the user asks to discover, inspect, summarize, understand, brief, or orient around a Bloomreach/Loomi project before recommending CRM segments, scenarios, lifecycle opportunities, or content.
---

# Loomi Project Discovery Brief

Use this skill to turn raw Loomi MCP context into a concise strategic brief that downstream CRM tools can use.

## Workflow

1. Use `docs/agent-context/router.md` to classify the task.
2. Load relevant local context from `contexts/`.
3. If live context is needed or local workspace context is stale, discover project context with Loomi MCP.
   - Prefer: event schema, customer properties, existing segmentations, customer filters, scenarios, campaigns, recommendations, predictions, consent settings, and frequency policies.
   - Do not fetch customer PII unless the user explicitly asks and the task truly requires it.
4. Read local business context only when relevant:
   - `contexts/team.md`
   - `contexts/industries/<industry>.md`
   - relevant advisor prompts under `prompts/advisors/`
5. If Bloomreach CLI workspace context is supplied, use selected cards and raw fallback only when they add project evidence.
6. Build a task packet with goal, exact IDs, loaded facts, assumptions, scope, and missing context.
7. Call `build_project_brief`.
8. Return the brief in human form, then keep the structured JSON available for the next tool call.

## Tool Input

Pass the selected project evidence from the task packet as a stringified JSON object in `projectContextJson`. If the evidence came from Loomi MCP discovery, keep the same field name for tool compatibility.

```json
{
  "request": "Discover the project and find lifecycle opportunities",
  "projectId": "<project uuid or empty>",
  "industry": "ecommerce",
  "teamContext": "<team.md text or empty>",
  "industryContext": "<industry pack text or empty>",
  "projectContextJson": "{\"scenarios\":[],\"event_schema\":[]}"
}
```

## Output Standard

Always summarize:

- data strengths
- data gaps
- lifecycle coverage
- available channels
- consent and frequency constraints
- opportunity themes
- recommended next tool call

Do not recommend specific scenarios until the brief identifies the usable data and constraints.