> ## Documentation Index
> Fetch the complete documentation index at: https://docs.evermuse.com/llms.txt
> Use this file to discover all available pages before exploring further.

# Skills & prompts

> How Evermuse skills give your agent a product-research methodology, and how to run them as MCP prompts and slash commands.

**Skills** are Evermuse's library of product-management and research methodologies: step-by-step guides for writing
a PRD, analysing competitors, building personas, sizing a market and more. Each skill tells the agent how to gather
customer evidence with the Evermuse tools, how to weigh it, and what a good result looks like. That's what keeps
answers grounded and cited rather than generic.

Your agent can use skills in three ways:

<CardGroup cols={3}>
  <Card title="Automatically" icon="sparkles">
    Workflow tools and search tools attach the relevant skill to their results when `include_guidance` is `true` (the
    default).
  </Card>

  <Card title="On demand" icon="magnifying-glass">
    The agent finds skills with `find_skills` and loads them with `read_skills`. Both are free.
  </Card>

  <Card title="As prompts" icon="terminal">
    Your client can list skills as MCP prompts, which many clients show as slash commands.
  </Card>
</CardGroup>

## Prompts

The server exposes every enabled skill in your workspace through `prompts/list`. Each prompt:

* is named after the skill id, for example `create-prd` or `competitor-analysis` (custom skills use their Evermuse
  skill id);
* has the skill's title and description;
* takes one optional argument, `input`: what to run the skill on (a product, feature, competitor or question).

`prompts/get` returns the skill as a single user message: Evermuse's citation rules, then the full skill and its
references, with your `input` filled in.

How prompts appear depends on your client. For example, Claude Code shows them as `/mcp__evermuse__create-prd`, and
Claude shows them in the attachment menu under the Evermuse connector.

<Tip>
  Skills that your workspace admin has disabled don't appear as prompts. Custom skills your team adds in Evermuse appear
  alongside the built-in ones.
</Tip>

## Built-in skills

<AccordionGroup>
  <Accordion title="Customer research & discovery">
    | Skill | What it does |
    | - | - |
    | `customer-research` | Research what customers think, need, ask for or complain about. |
    | `deep-dive` | A quote-rich, source-linked brief on one topic you care about. |
    | `analyze-feature-requests` | Triage feature requests into themes, merge duplicates and surface the need behind each ask. |
    | `sentiment-analysis` | Sentiment, themes and satisfaction shifts over a time period, with a score per theme. |
    | `summarize-conversation` | Summarize one meeting, call, thread or document. |
    | `daily-brief` | A recurring brief of what Evermuse learned from customer sources recently. |
    | `interview-script` | A Mom-Test interview guide that skips what customers have already answered. |
    | `customer-journey-map` | The journey from awareness to advocacy, with a real pain point and customer quote at each stage. |
    | `user-personas` | Build evidence-backed user personas. |
    | `ideal-customer-profile` | An ideal customer profile built from your real, happy customers. |
    | `segmentation` | Segment a market or user base into 3–5 groups by evidenced need differences. |
  </Accordion>

  <Accordion title="Product definition">
    | Skill | What it does |
    | - | - |
    | `create-prd` | A business PRD grounded in customer evidence. |
    | `write-feature-spec` | A testable feature specification. |
    | `user-stories` | INVEST user stories with evidence-backed acceptance criteria. |
    | `test-scenarios` | Given/When/Then QA scenarios from acceptance criteria and real customer edge cases. |
    | `development-plan` | A phased development plan that starts from what customers requested. |
    | `assumptions` | Surface and score the risky assumptions behind an idea, checking the evidence first. |
    | `brainstorm-ideas` | Product ideas from PM, design and engineering perspectives, seeded by real unmet needs. |
    | `release-notes` | User-facing release notes with a "You asked, we built" angle. |
    | `review-pr` | Review a pull request against code quality, its spec and the customer asks behind it. |
    | `shipping-artifacts` | Generate ship-readiness documentation for an AI-built feature. |
  </Accordion>

  <Accordion title="Strategy & prioritization">
    | Skill | What it does |
    | - | - |
    | `prioritize-features` | Rank a backlog using real demand counts from customer evidence. |
    | `opportunity-solution-tree` | An opportunity solution tree where every opportunity is backed by evidence. |
    | `outcome-roadmap` | Turn a feature-list roadmap into an outcome-focused one. |
    | `product-strategy` | A Product Strategy Canvas anchored in customer evidence and market context. |
    | `product-vision` | A product vision anchored in customers' own words. |
    | `north-star-metric` | A customer-centric North Star Metric with 3–5 input metrics. |
    | `brainstorm-okrs` | Team OKRs where each objective maps to an evidenced customer outcome. |
    | `strategy-frameworks` | SWOT, PESTLE, Porter's Five Forces or Ansoff, with every cell cited. |
    | `strategy-red-team` | Attack a strategy, PRD or roadmap with the customer counter-evidence. |
    | `gap-analysis` | Audit a spec, plan or shipped feature for gaps against other specs, the code and customer needs. |
  </Accordion>

  <Accordion title="Market & go-to-market">
    | Skill | What it does |
    | - | - |
    | `competitor-analysis` | Analyze competitors and find differentiation openings. |
    | `competitive-battlecard` | A sales battlecard built from real objections and competitor mentions. |
    | `market-sizing` | TAM, SAM and SOM, top-down and bottom-up, validated against customer evidence. |
    | `beachhead-segment` | Pick the beachhead segment that shows up most, and most painfully, in the evidence. |
    | `positioning-and-messaging` | Positioning options and messages written in your customers' own vocabulary. |
    | `value-proposition` | A jobs-to-be-done value proposition plus ready-to-use marketing, sales and onboarding statements. |
    | `gtm-strategy` | Messaging, channels, segments, motions, metrics and a launch timeline. |
    | `pricing-strategy` | Pricing grounded in willingness-to-pay signals and pricing objections. |
    | `business-model` | A Business Model Canvas with evidence-grounded problem and segment blocks. |
  </Accordion>

  <Accordion title="Using Evermuse">
    | Skill | What it does |
    | - | - |
    | `using-evermuse` | The core rules for searching Evermuse, weighing evidence and citing it. Loaded by every other skill. |
    | `setup-evermuse` | Guided setup of Evermuse inside your AI assistant. |
    | `schedule-new-ingestion-tasks` | Schedule recurring ingestion from newly connected customer-feedback sources in Claude or ChatGPT. |
    | `list-capabilities` | Show everything Evermuse can do. |
    | `initial-report` | Present the first Evermuse report after your sources are scanned. |
  </Accordion>
</AccordionGroup>

## Guidance in tool results

With `include_guidance: true`, these tools attach a skill to their result:

| Tool | Skill attached |
| - | - |
| Each [workflow tool](/mcp/tools#workflow-tools) | The matching skill, for example `create-prd` for `create_prd` (`write_brief` uses `daily-brief`) |
| `search`, `find_sources` | `using-evermuse` |
| `get_opportunities` | `prioritize-features` |
| `list_competitors`, `get_competitor_capabilities` | `competitor-analysis` |
| `find_skills` | The best-matching skill, plus your products and the `using-evermuse` doctrine |

Guidance is optional and never blocks a call: if a skill can't be loaded, the tool returns its data without it. Once
the agent has the methodology, it can pass `include_guidance: false` to fit more data into each page.


This documentation is built and hosted on [Mintlify](https://mintlify.com), a developer documentation platform.