Skip to content
Skills Library
Management & product · Analysis

brief

/brief

The front door to PM Pilot. Turns one decision question into an evidence-backed decision brief where every claim carries its source, counter-evidence is stated, and missing evidence is named rather than papered over. Works on a local folder with no MCP setup. Triggers on: "brief", "/brief", "should we", "decision brief", "help me decide", "make the case for", "what does the evidence say about".

Repository
pm-pilot ↗
Licence
MIT
Last repository change
02.10.2026

Assessment and security

100/ 100Class: good
Format25 / 25
Frontmatter with name and description, valid name, description up to 1,024 characters, name matching the folder.
Discoverability25 / 25
Description explains when to use the skill, has a reasonable length and uses third person.
Content35 / 35
At least 150 words and at most 500 lines, headings, step lists, examples or code.
Hygiene15 / 15
No placeholders such as TODO, no Windows paths, with a licence.
Riskno signals detected in the scan
Files in this package
1
Scripts
0
HTML comments
0
URLs in SKILL.md
0
Deprecation

No signals detected. Repository changed on 02.10.2026, without references to retired models or APIs.

Licence

MIT, open: use and modify while retaining author attribution.

What it does and when to use it

The front door to PM Pilot. Turns one decision question into an evidence-backed decision brief where every claim carries its source, counter-evidence is stated, and missing evidence is named rather than papered over. Works on a local folder with no MCP setup. Triggers on: "brief", "/brief", "should we", "decision brief", "help me decide", "make the case for", "what does the evidence say about".

Triggers from the description

briefshould wedecision briefhelp me decidemake the case forwhat does the evidence say about
Original author description

The front door to PM Pilot. Turns one decision question into an evidence-backed decision brief where every claim carries its source, counter-evidence is stated, and missing evidence is named rather than papered over. Works on a local folder with no MCP setup. Triggers on: "brief", "/brief", "should we", "decision brief", "help me decide", "make the case for", "what does the evidence say about".

How it works

  1. When to Activate
  2. The contract
  3. Execution
  4. Output format
  5. Rules
  6. Integration with other skills
Original SKILL.md headings · 145 lines, 1,066 words
SKILL.md preview
---
name: brief
description: |
  The front door to PM Pilot. Turns one decision question into an evidence-backed decision brief where every claim carries its source, counter-evidence is stated, and missing evidence is named rather than papered over. Works on a local folder with no MCP setup. Triggers on: "brief", "/brief", "should we", "decision brief", "help me decide", "make the case for", "what does the evidence say about".
origin: pm-pilot
version: 1.0.0
---

# Brief

One command, one decision, one reviewable artifact.

```
/pm-core:brief "Should we prioritize bulk import for mid-market?"
```

This is the first thing a new user should run. It needs no MCP servers, no configuration, and no knowledge of the other skills. It reads a folder of evidence, produces a decision brief, and forces a decision state at the end.

## When to Activate

- User runs `/pm-core:brief` with or without a question
- User asks "should we X", "help me decide on X", "make the case for X"
- User asks what the evidence says about a decision they are facing
- A new user has just installed PM Pilot and does not know where to start

## The contract

A brief is not a summary. It is an argument with its sources attached, and it is only useful if a reader can check it. So:

- Every consequential claim names where it came from.
- Every direct quote is verified against its source before delivery.
- Evidence that cuts against the recommendation is stated, not omitted.
- Evidence that does not exist is named as missing, not inferred.
- The brief ends in a decision state, not in a draft.

If you cannot meet all five, say which one failed and why. Never deliver a brief that looks complete when it is not.

## Execution

### Step 1: Name the decision

Take the decision from the user's argument. If they ran the command bare, ask once:

> What decision are you trying to make? One sentence, phrased as a question.

Reject a topic disguised as a decision. "Bulk import" is a topic. "Should we prioritize bulk import for mid-market this quarter?" is a decision. If the user gives a topic, offer the sharpest decision question you can infer from it and ask them to confirm or correct it in one line.

Do not proceed past a vague decision. Everything downstream inherits the framing.

### Step 2: Confirm sources, once

Look for evidence in this order and show the user what you found:

1. The current folder and any `evidence/`, `research/`, `interviews/`, `notes/`, `transcripts/` subfolder
2. `context/product.md`, `context/company.md`, `context/personas.md`, `context/competitors.md` if they have content
3. Connected MCP tools, if any are available (Jira, Slack, Confluence, Granola, GitHub)

Report it as a short list and ask one question:

> Found 14 files in `research/` and 3 context files. Jira and Granola are connected. Use all of these, or point me somewhere else?

Accept a yes and move. Accept a path and use it. This is the only confirmation gate in the whole flow. Do not ask again later.

If nothing is found, say so plainly and ask for a folder or a paste. Never invent a source, and never produce a brief from the model's own background knowledge while implying it came from the user's evidence.

### Step 3: Gather and verify

Route to the skills that already exist rather than duplicating them:

| Evidence type | Skill to use |
|---|---|
| Customer interviews and transcripts | `/pm-discovery:interview-snapshot`, then `/pm-core:synthesize-interviews` |
| Cross-tool history on a topic | `/pm-core:deep-context` |
| Competitive and market claims | `/pm-content:market-research` |
| A messy causal question underneath the decision | `/pm-core:issue-tree` |

Apply the hallucination guard from `synthesize-interviews` to every direct quote regardless of which path produced it. Search the source for at least 60% of the quote's words, exact match, case insensitive. Flag any miss inline with `[UNVERIFIED - edit before citing]` and keep going. Never silently repair a quote.

### Step 4: Write the brief

Use the output format below. Keep it to one screen for the body. Depth belongs in the evidence table, not in the prose.

### Step 5: Force a decision state

A brief that ends as a draft has not done its job. Close by asking the user to pick one:

- **Decide.** The evidence is sufficient. Record the call and the reasoning.
- **Defer.** Right call, wrong time. Record what has to change for this to come back.
- **Experiment.** The uncertainty is testable. Record the smallest test that would settle it.
- **Research.** The evidence gap is real. Record exactly what is missing and who can get it.

Write their answer into the brief before saving. A brief with no decision state is incomplete.

## Output format

```markdown
# Decision Brief: <the decision question>

**Date:** YYYY-MM-DD
**Sources reviewed:** <n files, n tools>
**Quotes verified:** 12/13 verified against source

## Recommendation

<Two sentences. The call, and the single strongest reason for it.>

## The case for

- <Claim.> [source: interviews/acme-2026-08-14.md, verified]
- <Claim.> [source: Jira INS-3128, verified]

## The case against

- <Claim that cuts against the recommendation.> [source: ...]

State at least one, and mean it. If the evidence genuinely runs one way, write
"No counter-evidence found in the reviewed sources" and treat that as a warning
sign about the sources rather than a strength of the argument.

## What we do not know

- <Missing evidence, and what it would take to get it.>
- <Question no source in this set can answer.>

## Confidence

<One of: high, medium, low.> <One sentence naming what would move it.>

## Decision

**State:** decide | defer | experiment | research
**Call:** <what was decided>
**Revisit when:** <trigger>
```

Save to `briefs/YYYY-MM-DD-<slug>.md` unless the user says otherwise.

## Rules

- One confirmation gate, in step 2. Everything else runs without asking.
- Never present an inference as a sourced claim. An inference is labelled as yours.
- A claim you cannot trace to a source in the confirmed set does not go in the brief. It goes in "What we do not know".
- Do not soften a weak evidence base with hedged prose. Say the base is weak and name what is missing.
- Web search is allowed for market and competitor claims, and those get their URL in the source tag like any other claim.
- If the user's own prior decisions contradict the recommendation, surface that in "The case against".

## Integration with other skills

- **Upstream**: nothing. This is the entry point.
- **Downstream**: `/pm-core:prd` turns a decided brief into a spec. `/pm-core:critique` pressure-tests the brief before it goes to a stakeholder. `/crucible:crucible` for a decision big enough to deserve a council.
- **Sibling**: `/pm-core:prioritize` when the question is a ranking across many items rather than one call.

Versions

RepositoryScoreLicence
This page · mshadmanrahman/pm-pilot100MIT
sekretmaksym-dev/design-skills100MIT
digital-stoic-org/agent-skills100MIT
wrg32786/aigent-os100MIT
magnifito/domain-name-engine100MIT
alirezarezvani/claude-skills100MIT
anthropics/knowledge-work-plugins100Apache-2.0
educlopez/ui-craft100MIT
oliver-kriska/claude-elixir-phoenix100MIT

Database snapshot: 13.11.2026 · package version a698ee559c93 (SHA-256 digest of the entire package)

Downloads in the last 30 days: 0 · ranking

Report a problem with this skill: harmful content, incorrect descriptions or an author’s removal request.

Similar skills

  • 90research-brief
    Research a question and write a short, sourced brief. Use when the user says "research", "look into", "find out", "what does the evidence say", "is it true that", or needs a fact base before a decision.
  • 100research-brief
    Use when gathering evidence before a decision, summarizing repo or external facts, checking whether context supports a proposed path, or preparing a brief before specs, plans, brainstorming, or implementation.
  • 100decision-question
    Ask the user to choose only when blocked on a decision they own, with the smallest complete decision brief. Use when you say `ask the user to choose`, `do I need a user decision here`, or `present a decision brief`. Do NOT use to review a decision already made (use…
  • 90decision-brief
    Prepares a decision brief: question, sourced facts, assumptions, options with trade-offs, recommendation and decision maker. Use when a leader faces a significant choice such as an investment, hire, price change or partnership. The brief decides nothing.
  • 100decision-question
    Ask the user to choose only when blocked on a decision they own, with the smallest complete decision brief. Use when you say `ask the user to choose`, `do I need a user decision here`, or `present a decision brief`. Do NOT use to review a decision already made (use…
  • 95create-brief
    Use this skill when the user says 'I want to build', 'new app idea', 'create a brief', 'product brief', 'help me define what I am building', or when no existing brief or PRD exists in the docs/ folder. This skill translates a vague idea into a structured Product Brief. It asks 5…

Feedback and reports

Your message goes to Bartosz Sokoliński. An email address is needed only for a reply. See the privacy policy.