silverstein/minutes

## Skill Path

Before running helper scripts or opening bundled references, set:

```bash
export MINUTES_SKILLS_ROOT="$(git rev-parse --show-toplevel)/.opencode/skills"
export MINUTES_SKILL_ROOT="$MINUTES_SKILLS_ROOT/minutes-brief"
```

# /minutes-brief

Fast, non-interactive briefing that synthesizes your relationship history with someone into a one-page brief — designed so you can read it in 60 seconds before walking into a call.

This is the **proactive layer**. It's built to be invoked silently by a hook (e.g., 15 min before a calendar event) and to also work as a manual `/minutes-brief` command. Unlike `/minutes-prep`, brief asks no questions and sets no goals — it just hands you the facts.

# /minutes-cleanup

Help the user manage disk space and organize old recordings. Minutes is
transcript-first: markdown notes and structured memory are durable, while raw
audio is a temporary recovery/reprocessing layer unless pinned.

## Check current usage

```bash
minutes storage
minutes storage --json
```

Present this to the user before taking any action.

# /minutes-copilot

Use the real Minutes copilot runtime to start or control Coach. Do not build a transcript reader, event poller, prompt loop, or shell tailer in this skill.

If the user explicitly asks **you or the current terminal agent** to watch, advise, or strategize, use `/minutes-live-sidekick` instead. If the user says only "coach me live" or otherwise leaves the surface ambiguous, ask exactly: "Do you want me in this terminal to be your sidekick, or should I open the Minutes Coach HUD?" Do not start Coach while clarifying.

## Start Coach

1. Get one concrete meeting goal. Use the user's stated goal; if none is present, ask exactly: "What outcome should Coach help you achieve in this meeting?"
2. Choose a supported mode only when the user makes it clear: `sales`, `discovery`, `interview`, `negotiation`, `difficult-conversation`, `decision`, or `generic`. Default to `generic`.
3. If first-party Minutes copilot MCP controls are actually available, call their start control with the same goal, mode, and `tui` surface. Never invent an MCP tool name.
4. Otherwise invoke the foreground CLI, passing the goal as one safely escaped argument:

```bash

## Skill Path

Before running helper scripts or opening bundled references, set:

```bash
export MINUTES_SKILLS_ROOT="$(git rev-parse --show-toplevel)/.opencode/skills"
export MINUTES_SKILL_ROOT="$MINUTES_SKILLS_ROOT/minutes-debrief"
```

# /minutes-debrief

Post-meeting analysis that reads your latest recording, compares what happened to what you planned, and surfaces decision evolution — so nothing falls through the cracks.

## How it works

# /minutes-graph

Minutes builds relationship rankings, exact person profiles, and commitments from one supervised, process-private SQLite projection of stable policy-authorized Markdown plus confirmed identity corrections. One ordered snapshot authority spans corpus and corrections, and the worker is hard-limited for memory, output, and wall time. Topic research uses the separately bounded live-source search boundary. Both paths re-attest policy before returning facts. Do not fall back to a retired durable index or read meeting files directly.

## Privacy boundary

- Never walk meeting files, parse frontmatter yourself, or read raw transcripts for this skill.
- Never run `graph_build.py` or read `~/.minutes/graph/index.json`; those are retired legacy surfaces.
- Never create a replacement graph cache, spreadsheet, JSON file, or database.
- Never pass `--include-restricted`. Restricted meetings are intentionally absent from this agent-facing skill.
- Treat any authorization, resource-budget, correction-race, or projection error as a hard stop. Do not fall back to filesystem reads.

## Available commands

# /minutes-ideas — Recent Voice Memos & Ideas

Surface voice memos and ideas captured from any device in the last 14 days.
This is the recall layer for the cross-device ghost context pipeline.

## How to run

1. Search for recent voice memos using the `minutes` CLI:

```bash
minutes list --content-type memo --limit 20
```

2. Require exit status 0 and use the JSON written to stdout. If the CLI is

# /minutes-ingest

Process meetings through the knowledge extraction pipeline to update person profiles, append to the knowledge log, and maintain the index.

## Prerequisites

The `[knowledge]` section must be configured in `~/.config/minutes/config.toml`:

```toml
[knowledge]
enabled = true
path = "/path/to/knowledge/base"
adapter = "wiki"  # or "para", "obsidian"
engine = "none"   # or "agent" for LLM extraction

# /minutes-lint

Run a consistency check across policy-authorized normal meetings to find
decision conflicts and stale commitments. Restricted meetings are excluded from
this agent workflow by default and must never be described as absent evidence.

## How to run the lint

1. **Run the consistency check**:
   ```bash
   minutes consistency --stale-after-days 14
   ```

   Optional filters:

# /minutes-list

Show recent meetings and voice memos, sorted newest-first.

## Usage

```bash
# List last 10 recordings (default)
minutes list

# Show more
minutes list --limit 20

# Only voice memos

# /minutes-live-sidekick

Act as the live meeting sidekick in the current terminal agent session. Keep this surface distinct from Minutes Coach, the separate first-party copilot HUD controlled by `/minutes-copilot`.

## Select the surface

- If the user explicitly asks **you or the terminal agent** to watch, advise, or strategize, continue with this skill.
- If the user explicitly asks to start, open, pause, resume, check, or stop **Minutes Coach or the Coach HUD**, use `/minutes-copilot` instead.
- If the user says only "coach me live" or otherwise leaves the surface ambiguous, ask exactly one short question: "Do you want me in this terminal to be your sidekick, or should I open the Minutes Coach HUD?"

Do not start Coach while clarifying.

## Establish the contract

## Skill Path

Before running helper scripts or opening bundled references, set:

```bash
export MINUTES_SKILLS_ROOT="$(git rev-parse --show-toplevel)/.opencode/skills"
export MINUTES_SKILL_ROOT="$MINUTES_SKILLS_ROOT/minutes-mcp-recall"
```

# /minutes-mcp-recall

Route a meeting-memory question to the smallest useful Minutes MCP tool. This
skill is tool-name-oriented for MCP hosts such as OpenClaw, Hermes-agent, and
other MCP clients. Use `minutes-search` instead when the host only has CLI or

## Skill Path

Before running helper scripts or opening bundled references, set:

```bash
export MINUTES_SKILLS_ROOT="$(git rev-parse --show-toplevel)/.opencode/skills"
export MINUTES_SKILL_ROOT="$MINUTES_SKILLS_ROOT/minutes-mirror"
```

# /minutes-mirror

Self-coaching analysis based on your own meeting transcripts. Two modes:

- **Single-meeting mode** — review a specific meeting and surface what you did, what was unusual for you, and one concrete thing to try next time.

# /minutes-note

Add a timestamped note during a recording, or annotate a past meeting.

## During a recording

```bash
# Add a note to the active recording (auto-timestamped)
minutes note "Alex wants monthly billing not annual billing"
minutes note "Case agreed — compromise at monthly billing for experiment"
```

Each note gets a timestamp matching the recording position (e.g., `[4:23]`). Notes feed into the LLM summarizer as high-priority context — the AI knows what you thought was important and weights those parts of the transcript more heavily in the summary.

## Skill Path

Before running helper scripts or opening bundled references, set:

```bash
export MINUTES_SKILLS_ROOT="$(git rev-parse --show-toplevel)/.opencode/skills"
export MINUTES_SKILL_ROOT="$MINUTES_SKILLS_ROOT/minutes-prep"
```

# /minutes-prep

Interactive meeting preparation that searches the normal conversation history
authorized by Minutes, synthesizes a relationship brief, and produces talking
points — before you walk into the room. Restricted history is excluded from

## Skill Path

Before running helper scripts or opening bundled references, set:

```bash
export MINUTES_SKILLS_ROOT="$(git rev-parse --show-toplevel)/.opencode/skills"
export MINUTES_SKILL_ROOT="$MINUTES_SKILLS_ROOT/minutes-recap"
```

# /minutes-recap

Synthesize today's policy-authorized normal meetings and voice memos into a
single daily brief. Restricted history is excluded from this agent workflow by
default and must never be treated as absent evidence.

## Skill Path

Before running helper scripts or opening bundled references, set:

```bash
export MINUTES_SKILLS_ROOT="$(git rev-parse --show-toplevel)/.opencode/skills"
export MINUTES_SKILL_ROOT="$MINUTES_SKILLS_ROOT/minutes-record"
```

# /minutes-record

Record audio from the microphone, transcribe it locally with whisper.cpp, and save as searchable markdown.

## How it works

# /minutes-release-notes

Draft release notes that explain why a Minutes release matters without making users decode the commit history. Produce a draft only. Do not create, edit, or publish a GitHub release unless the user explicitly asks.

## Inputs

Collect:

- the new version, such as `v0.22.0`
- the previous stable tag, or an explicit starting ref
- the target ref, normally `HEAD`
- the release channel, normally stable or preview
- any known breaking changes, migrations, compatibility notes, or contributor credits

# /minutes-search

Find information across all meeting transcripts and voice memos.

## Usage

```bash
# Basic search
minutes search "pricing strategy"

# Filter to just voice memos
minutes search "onboarding idea" -t memo

# Filter to just meetings

## Skill Path

Before running helper scripts or opening bundled references, set:

```bash
export MINUTES_SKILLS_ROOT="$(git rev-parse --show-toplevel)/.opencode/skills"
export MINUTES_SKILL_ROOT="$MINUTES_SKILLS_ROOT/minutes-setup"
```

# /minutes-setup

Walk the user through first-time Minutes setup, step by step.

## Setup steps

## Skill Path

Before running helper scripts or opening bundled references, set:

```bash
export MINUTES_SKILLS_ROOT="$(git rev-parse --show-toplevel)/.opencode/skills"
export MINUTES_SKILL_ROOT="$MINUTES_SKILLS_ROOT/minutes-tag"
```

# /minutes-tag

Lightweight outcome tagging — adds an `outcome:` field to a meeting's frontmatter so `/minutes-mirror` can correlate the user's behavior with their results over time.

The whole point of this skill is **speed**. Tagging should take 5 seconds, not 5 questions. Don't be precious about it — most users will never adopt tagging if it feels like data entry.

## Skill Path

Before running helper scripts or opening bundled references, set:

```bash
export MINUTES_SKILLS_ROOT="$(git rev-parse --show-toplevel)/.opencode/skills"
export MINUTES_SKILL_ROOT="$MINUTES_SKILLS_ROOT/minutes-verify"
```

# /minutes-verify

Run a health check on the Minutes installation to confirm everything is working.

## How to verify

## Skill Path

Before running helper scripts or opening bundled references, set:

```bash
export MINUTES_SKILLS_ROOT="$(git rev-parse --show-toplevel)/.opencode/skills"
export MINUTES_SKILL_ROOT="$MINUTES_SKILLS_ROOT/minutes-video-review"
```

# /minutes-video-review

Analyze a product walkthrough, bug report video, Loom, ScreenPal, or local recording into a durable artifact bundle that agents can keep working from.

This skill is for **meeting-adjacent product artifacts**, not for generic "understand any video" requests. Use it when the user wants a recorded demo, bug repro, or walkthrough turned into something actionable for engineering, product, support, or follow-up agent work.

## Skill Path

Before running helper scripts or opening bundled references, set:

```bash
export MINUTES_SKILLS_ROOT="$(git rev-parse --show-toplevel)/.opencode/skills"
export MINUTES_SKILL_ROOT="$MINUTES_SKILLS_ROOT/minutes-weekly"
```

# /minutes-weekly

Synthesize a week of policy-authorized normal meetings and voice memos into a
forward-looking brief — themes, decision arcs, stale commitments, and what
deserves attention Monday. Restricted history is excluded by default.

# /minutes-x1-closeout

Turn one clear meeting outcome into a proposed closeout for one existing X1
coordination thread. Minutes supplies the source. X1 supplies identity,
authority, review, disposition, and the household record.

## Required connections

This workflow needs both the local Minutes MCP and the official X1 MCP. If X1
isn't connected, stop after identifying the meeting outcome and direct the user
to X1's official connection instructions. Don't invent an endpoint, credential,
account, thread ID, or local substitute.

## Workflow

# /minutes-x1-send-meeting

Send one of the user's own meetings to their X1 household for review. Minutes
supplies the meeting's outcomes. X1 decides who the sender is, which
household it goes to, and what gets matched, and the user approves the send
in X1. The meeting lands in the user's own review queue. Nothing reaches the
household record until they confirm each item, and no professional sees it.

## Required connections

This workflow needs both the local Minutes MCP and the official X1 MCP. Call
X1 `get_user_capabilities` first. If `submit_my_meeting` or
`request_human_confirmation` isn't mounted, stop: X1 hasn't turned meeting
sends on for this account yet. Don't substitute another X1 write tool, invent

# /minutes-brief

Fast, non-interactive briefing that synthesizes your relationship history with someone into a one-page brief — designed so you can read it in 60 seconds before walking into a call.

This is the **proactive layer**. It's built to be invoked silently by a hook (e.g., 15 min before a calendar event) and to also work as a manual `/minutes-brief` command. Unlike `/minutes-prep`, brief asks no questions and sets no goals — it just hands you the facts.

## How it works

### Phase 0: Determine the target

Three ways the user can invoke this:

**1. With a name** (`/minutes-brief sarah`, "brief me on Alex")
→ Use that name directly. Before searching, check for learned aliases:

# /minutes-cleanup

Help the user manage disk space and organize old recordings. Minutes is
transcript-first: markdown notes and structured memory are durable, while raw
audio is a temporary recovery/reprocessing layer unless pinned.

## Check current usage

```bash
minutes storage
minutes storage --json
```

Present this to the user before taking any action.

# /minutes-copilot

Use the real Minutes copilot runtime to start or control Coach. Do not build a transcript reader, event poller, prompt loop, or shell tailer in this skill.

If the user explicitly asks **you or the current terminal agent** to watch, advise, or strategize, use `/minutes-live-sidekick` instead. If the user says only "coach me live" or otherwise leaves the surface ambiguous, ask exactly: "Do you want me in this terminal to be your sidekick, or should I open the Minutes Coach HUD?" Do not start Coach while clarifying.

## Start Coach

1. Get one concrete meeting goal. Use the user's stated goal; if none is present, ask exactly: "What outcome should Coach help you achieve in this meeting?"
2. Choose a supported mode only when the user makes it clear: `sales`, `discovery`, `interview`, `negotiation`, `difficult-conversation`, `decision`, or `generic`. Default to `generic`.
3. If first-party Minutes copilot MCP controls are actually available, call their start control with the same goal, mode, and `tui` surface. Never invent an MCP tool name.
4. Otherwise invoke the foreground CLI, passing the goal as one safely escaped argument:

```bash

# /minutes-debrief

Post-meeting analysis that reads your latest recording, compares what happened to what you planned, and surfaces decision evolution — so nothing falls through the cracks.

## How it works

This is a multi-phase interactive flow. It connects to `/minutes-prep` when a prep file exists, creating a before→after loop.

### Phase 1: Find the most recent recording

```bash
minutes list --limit 5
```

# /minutes-graph

Minutes builds relationship rankings, exact person profiles, and commitments from one supervised, process-private SQLite projection of stable policy-authorized Markdown plus confirmed identity corrections. One ordered snapshot authority spans corpus and corrections, and the worker is hard-limited for memory, output, and wall time. Topic research uses the separately bounded live-source search boundary. Both paths re-attest policy before returning facts. Do not fall back to a retired durable index or read meeting files directly.

## Privacy boundary

- Never walk meeting files, parse frontmatter yourself, or read raw transcripts for this skill.
- Never run `graph_build.py` or read `~/.minutes/graph/index.json`; those are retired legacy surfaces.
- Never create a replacement graph cache, spreadsheet, JSON file, or database.
- Never pass `--include-restricted`. Restricted meetings are intentionally absent from this agent-facing skill.
- Treat any authorization, resource-budget, correction-race, or projection error as a hard stop. Do not fall back to filesystem reads.

## Available commands

# /minutes-ideas — Recent Voice Memos & Ideas

Surface voice memos and ideas captured from any device in the last 14 days.
This is the recall layer for the cross-device ghost context pipeline.

## How to run

1. Search for recent voice memos using the `minutes` CLI:

```bash
minutes list --content-type memo --limit 20
```

2. Require exit status 0 and use the JSON written to stdout. If the CLI is

# /minutes-ingest

Process meetings through the knowledge extraction pipeline to update person profiles, append to the knowledge log, and maintain the index.

## Prerequisites

The `[knowledge]` section must be configured in `~/.config/minutes/config.toml`:

```toml
[knowledge]
enabled = true
path = "/path/to/knowledge/base"
adapter = "wiki"  # or "para", "obsidian"
engine = "none"   # or "agent" for LLM extraction

# /minutes-lint

Run a consistency check across policy-authorized normal meetings to find
decision conflicts and stale commitments. Restricted meetings are excluded from
this agent workflow by default and must never be described as absent evidence.

## How to run the lint

1. **Run the consistency check**:
   ```bash
   minutes consistency --stale-after-days 14
   ```

   Optional filters:

# /minutes-list

Show recent meetings and voice memos, sorted newest-first.

## Usage

```bash
# List last 10 recordings (default)
minutes list

# Show more
minutes list --limit 20

# Only voice memos

# /minutes-live-sidekick

Act as the live meeting sidekick in the current terminal agent session. Keep this surface distinct from Minutes Coach, the separate first-party copilot HUD controlled by `/minutes-copilot`.

## Select the surface

- If the user explicitly asks **you or the terminal agent** to watch, advise, or strategize, continue with this skill.
- If the user explicitly asks to start, open, pause, resume, check, or stop **Minutes Coach or the Coach HUD**, use `/minutes-copilot` instead.
- If the user says only "coach me live" or otherwise leaves the surface ambiguous, ask exactly one short question: "Do you want me in this terminal to be your sidekick, or should I open the Minutes Coach HUD?"

Do not start Coach while clarifying.

## Establish the contract

# /minutes-mcp-recall

Route a meeting-memory question to the smallest useful Minutes MCP tool. This
skill is tool-name-oriented for MCP hosts such as OpenClaw, Hermes-agent, and
other MCP clients. Use `minutes-search` instead when the host only has CLI or
shell access.

Read [the complete tool map](references/tool-map.md) only when the decision
table does not cover the request.

## Decision table

| User is asking... | Call |
|---|---|

# /minutes-mirror

Self-coaching analysis based on your own meeting transcripts. Two modes:

- **Single-meeting mode** — review a specific meeting and surface what you did, what was unusual for you, and one concrete thing to try next time.
- **Pattern mode** — surface trends across the last 30 days, including (if meetings are tagged) what behaviors correlate with winning vs losing.

The point is not to roast you. The point is to give you a kind, evidence-based mirror to behaviors that are usually invisible to you because you're inside them.

## How it works

### Phase 0: Identify "you"

Mirror needs to know which speaker label in the transcript is the user. Real transcripts use one of two formats:

# /minutes-note

Add a timestamped note during a recording, or annotate a past meeting.

## During a recording

```bash
# Add a note to the active recording (auto-timestamped)
minutes note "Alex wants monthly billing not annual billing"
minutes note "Case agreed — compromise at monthly billing for experiment"
```

Each note gets a timestamp matching the recording position (e.g., `[4:23]`). Notes feed into the LLM summarizer as high-priority context — the AI knows what you thought was important and weights those parts of the transcript more heavily in the summary.

# /minutes-prep

Interactive meeting preparation that searches the normal conversation history
authorized by Minutes, synthesizes a relationship brief, and produces talking
points — before you walk into the room. Restricted history is excluded from
this agent workflow by default and must never be treated as absent evidence.

## How it works

This is a multi-phase interactive flow, not a single command. Walk the user through each phase using AskUserQuestion, pushing back on vague answers.

### Phase 0: Calendar auto-detect (optional, runs first)

Before asking who the user is meeting with, check if upcoming meetings are available from any calendar source. Try these in order — use the first that works:

# /minutes-recap

Synthesize today's policy-authorized normal meetings and voice memos into a
single daily brief. Restricted history is excluded from this agent workflow by
default and must never be treated as absent evidence.

## How to generate the recap

1. **Find today's recordings** from the bounded live list:
   ```bash
   minutes list --limit 50
   ```
   Require exit status 0, parse the JSON written to stdout, and select entries
   whose `date` is today. A date string is not a search query.

# /minutes-record

Record audio from the microphone, transcribe it locally with whisper.cpp, and save as searchable markdown.

## How it works

Recording is a two-step process — start and stop. Between those two commands, audio is captured continuously from the default input device.

**Start recording:**
```bash
minutes record
# Or with a title:
minutes record --title "Weekly standup with Alex"
```

# /minutes-release-notes

Draft release notes that explain why a Minutes release matters without making users decode the commit history. Produce a draft only. Do not create, edit, or publish a GitHub release unless the user explicitly asks.

## Inputs

Collect:

- the new version, such as `v0.22.0`
- the previous stable tag, or an explicit starting ref
- the target ref, normally `HEAD`
- the release channel, normally stable or preview
- any known breaking changes, migrations, compatibility notes, or contributor credits

# /minutes-search

Find information across all meeting transcripts and voice memos.

## Usage

```bash
# Basic search
minutes search "pricing strategy"

# Filter to just voice memos
minutes search "onboarding idea" -t memo

# Filter to just meetings

# /minutes-setup

Walk the user through first-time Minutes setup, step by step.

## Setup steps

### 1. Check current state first

Run the verify skill's script to see what's already done:
```bash
bash "${CLAUDE_PLUGIN_ROOT}/skills/minutes-verify/scripts/verify-setup.sh"
```

Skip any steps that already pass.

# /minutes-tag

Lightweight outcome tagging — adds an `outcome:` field to a meeting's frontmatter so `/minutes-mirror` can correlate the user's behavior with their results over time.

The whole point of this skill is **speed**. Tagging should take 5 seconds, not 5 questions. Don't be precious about it — most users will never adopt tagging if it feels like data entry.

## How it works

### Phase 1: Identify the meeting

Three patterns the user might use. **Always filter to meetings, not voice memos** — voice memos can't be "won" or "lost".

**1. Most recent** ("tag this meeting", "mark that as a win", "tag the call I just finished"):
```bash

# /minutes-verify

Run a health check on the Minutes installation to confirm everything is working.

## How to verify

Run the verification script included with this skill:

```bash
bash "${CLAUDE_PLUGIN_ROOT}/skills/minutes-verify/scripts/verify-setup.sh"
```

The script checks each component and outputs a pass/fail status for each. Read the output and report results to the user.

# /minutes-video-review

Analyze a product walkthrough, bug report video, Loom, ScreenPal, or local recording into a durable artifact bundle that agents can keep working from.

This skill is for **meeting-adjacent product artifacts**, not for generic "understand any video" requests. Use it when the user wants a recorded demo, bug repro, or walkthrough turned into something actionable for engineering, product, support, or follow-up agent work.

## What this skill does

The bundled script handles the deterministic pipeline:

- resolve a local file or hosted video URL
- download hosted video when needed
- extract audio with `ffmpeg`
- transcribe with Minutes first, using the user's existing Minutes transcription setup

# /minutes-weekly

Synthesize a week of policy-authorized normal meetings and voice memos into a
forward-looking brief — themes, decision arcs, stale commitments, and what
deserves attention Monday. Restricted history is excluded by default.

## How it works

This is a synthesis skill, not a command wrapper. It reads across the normal
meetings authorized by Minutes, cross-references decisions and action items,
and produces an intelligence brief. Restricted meetings are excluded from this
agent workflow by default and must never be treated as absent facts.

### Phase 1: Gather this week's recordings

# /minutes-x1-closeout

Turn one clear meeting outcome into a proposed closeout for one existing X1
coordination thread. Minutes supplies the source. X1 supplies identity,
authority, review, disposition, and the household record.

## Required connections

This workflow needs both the local Minutes MCP and the official X1 MCP. If X1
isn't connected, stop after identifying the meeting outcome and direct the user
to X1's official connection instructions. Don't invent an endpoint, credential,
account, thread ID, or local substitute.

## Workflow

# /minutes-x1-send-meeting

Send one of the user's own meetings to their X1 household for review. Minutes
supplies the meeting's outcomes. X1 decides who the sender is, which
household it goes to, and what gets matched, and the user approves the send
in X1. The meeting lands in the user's own review queue. Nothing reaches the
household record until they confirm each item, and no professional sees it.

## Required connections

This workflow needs both the local Minutes MCP and the official X1 MCP. Call
X1 `get_user_capabilities` first. If `submit_my_meeting` or
`request_human_confirmation` isn't mounted, stop: X1 hasn't turned meeting
sends on for this account yet. Don't substitute another X1 write tool, invent

## Skill Path

Before running helper scripts or opening bundled references, set:

```bash
export MINUTES_SKILLS_ROOT="$(git rev-parse --show-toplevel)/.agents/skills/minutes"
export MINUTES_SKILL_ROOT="$MINUTES_SKILLS_ROOT/minutes-brief"
```

# /minutes-brief

Fast, non-interactive briefing that synthesizes your relationship history with someone into a one-page brief — designed so you can read it in 60 seconds before walking into a call.

This is the **proactive layer**. It's built to be invoked silently by a hook (e.g., 15 min before a calendar event) and to also work as a manual `/minutes-brief` command. Unlike `/minutes-prep`, brief asks no questions and sets no goals — it just hands you the facts.

# /minutes-cleanup

Help the user manage disk space and organize old recordings. Minutes is
transcript-first: markdown notes and structured memory are durable, while raw
audio is a temporary recovery/reprocessing layer unless pinned.

## Check current usage

```bash
minutes storage
minutes storage --json
```

Present this to the user before taking any action.

# /minutes-copilot

Use the real Minutes copilot runtime to start or control Coach. Do not build a transcript reader, event poller, prompt loop, or shell tailer in this skill.

If the user explicitly asks **you or the current terminal agent** to watch, advise, or strategize, use `/minutes-live-sidekick` instead. If the user says only "coach me live" or otherwise leaves the surface ambiguous, ask exactly: "Do you want me in this terminal to be your sidekick, or should I open the Minutes Coach HUD?" Do not start Coach while clarifying.

## Start Coach

1. Get one concrete meeting goal. Use the user's stated goal; if none is present, ask exactly: "What outcome should Coach help you achieve in this meeting?"
2. Choose a supported mode only when the user makes it clear: `sales`, `discovery`, `interview`, `negotiation`, `difficult-conversation`, `decision`, or `generic`. Default to `generic`.
3. If first-party Minutes copilot MCP controls are actually available, call their start control with the same goal, mode, and `tui` surface. Never invent an MCP tool name.
4. Otherwise invoke the foreground CLI, passing the goal as one safely escaped argument:

```bash

## Skill Path

Before running helper scripts or opening bundled references, set:

```bash
export MINUTES_SKILLS_ROOT="$(git rev-parse --show-toplevel)/.agents/skills/minutes"
export MINUTES_SKILL_ROOT="$MINUTES_SKILLS_ROOT/minutes-debrief"
```

# /minutes-debrief

Post-meeting analysis that reads your latest recording, compares what happened to what you planned, and surfaces decision evolution — so nothing falls through the cracks.

## How it works

# /minutes-graph

Minutes builds relationship rankings, exact person profiles, and commitments from one supervised, process-private SQLite projection of stable policy-authorized Markdown plus confirmed identity corrections. One ordered snapshot authority spans corpus and corrections, and the worker is hard-limited for memory, output, and wall time. Topic research uses the separately bounded live-source search boundary. Both paths re-attest policy before returning facts. Do not fall back to a retired durable index or read meeting files directly.

## Privacy boundary

- Never walk meeting files, parse frontmatter yourself, or read raw transcripts for this skill.
- Never run `graph_build.py` or read `~/.minutes/graph/index.json`; those are retired legacy surfaces.
- Never create a replacement graph cache, spreadsheet, JSON file, or database.
- Never pass `--include-restricted`. Restricted meetings are intentionally absent from this agent-facing skill.
- Treat any authorization, resource-budget, correction-race, or projection error as a hard stop. Do not fall back to filesystem reads.

## Available commands

# /minutes-ideas — Recent Voice Memos & Ideas

Surface voice memos and ideas captured from any device in the last 14 days.
This is the recall layer for the cross-device ghost context pipeline.

## How to run

1. Search for recent voice memos using the `minutes` CLI:

```bash
minutes list --content-type memo --limit 20
```

2. Require exit status 0 and use the JSON written to stdout. If the CLI is

# /minutes-ingest

Process meetings through the knowledge extraction pipeline to update person profiles, append to the knowledge log, and maintain the index.

## Prerequisites

The `[knowledge]` section must be configured in `~/.config/minutes/config.toml`:

```toml
[knowledge]
enabled = true
path = "/path/to/knowledge/base"
adapter = "wiki"  # or "para", "obsidian"
engine = "none"   # or "agent" for LLM extraction

# /minutes-lint

Run a consistency check across policy-authorized normal meetings to find
decision conflicts and stale commitments. Restricted meetings are excluded from
this agent workflow by default and must never be described as absent evidence.

## How to run the lint

1. **Run the consistency check**:
   ```bash
   minutes consistency --stale-after-days 14
   ```

   Optional filters:

# /minutes-list

Show recent meetings and voice memos, sorted newest-first.

## Usage

```bash
# List last 10 recordings (default)
minutes list

# Show more
minutes list --limit 20

# Only voice memos

# /minutes-live-sidekick

Act as the live meeting sidekick in the current terminal agent session. Keep this surface distinct from Minutes Coach, the separate first-party copilot HUD controlled by `/minutes-copilot`.

## Select the surface

- If the user explicitly asks **you or the terminal agent** to watch, advise, or strategize, continue with this skill.
- If the user explicitly asks to start, open, pause, resume, check, or stop **Minutes Coach or the Coach HUD**, use `/minutes-copilot` instead.
- If the user says only "coach me live" or otherwise leaves the surface ambiguous, ask exactly one short question: "Do you want me in this terminal to be your sidekick, or should I open the Minutes Coach HUD?"

Do not start Coach while clarifying.

## Establish the contract

## Skill Path

Before running helper scripts or opening bundled references, set:

```bash
export MINUTES_SKILLS_ROOT="$(git rev-parse --show-toplevel)/.agents/skills/minutes"
export MINUTES_SKILL_ROOT="$MINUTES_SKILLS_ROOT/minutes-mcp-recall"
```

# /minutes-mcp-recall

Route a meeting-memory question to the smallest useful Minutes MCP tool. This
skill is tool-name-oriented for MCP hosts such as OpenClaw, Hermes-agent, and
other MCP clients. Use `minutes-search` instead when the host only has CLI or

## Skill Path

Before running helper scripts or opening bundled references, set:

```bash
export MINUTES_SKILLS_ROOT="$(git rev-parse --show-toplevel)/.agents/skills/minutes"
export MINUTES_SKILL_ROOT="$MINUTES_SKILLS_ROOT/minutes-mirror"
```

# /minutes-mirror

Self-coaching analysis based on your own meeting transcripts. Two modes:

- **Single-meeting mode** — review a specific meeting and surface what you did, what was unusual for you, and one concrete thing to try next time.

# /minutes-note

Add a timestamped note during a recording, or annotate a past meeting.

## During a recording

```bash
# Add a note to the active recording (auto-timestamped)
minutes note "Alex wants monthly billing not annual billing"
minutes note "Case agreed — compromise at monthly billing for experiment"
```

Each note gets a timestamp matching the recording position (e.g., `[4:23]`). Notes feed into the LLM summarizer as high-priority context — the AI knows what you thought was important and weights those parts of the transcript more heavily in the summary.