minutes-ingest
The minutes-ingest skill processes meeting notes through a structured extraction pipeline to populate a knowledge base with person profiles, chronological logs, and indexes. Use it when ingesting meeting files into wiki, PARA, or Obsidian systems, performing backfill operations on historical meetings, or syncing conversation data to maintain updated profile records. The tool extracts facts from YAML frontmatter with confidence filtering and deduplication safeguards, supporting both single-file and bulk processing modes with optional dry-run preview.
git clone --depth 1 https://github.com/silverstein/minutes /tmp/minutes-ingest && cp -r /tmp/minutes-ingest/tooling/skills/sources/minutes-ingest ~/.claude/skills/minutes-ingestskill.md
# /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
min_confidence = "strong"
```
If not configured, explain what's needed and offer to help set it up.
## How to run
### Single meeting
```bash
minutes ingest ~/meetings/2026-04-03-strategy-call.md
```
### All meetings (backfill)
```bash
minutes ingest --all
```
### Preview without writing (recommended first time)
```bash
minutes ingest --all --dry-run
```
## What it does
1. **Reads** each meeting's YAML frontmatter (decisions, action_items, entities, intents)
2. **Extracts** structured facts with confidence levels and source provenance
3. **Updates** person profiles in the knowledge base (adapter-dependent format)
4. **Appends** to `log.md` with a timestamped entry for each ingested meeting
5. **Skips** facts that already exist (deduplication) or are below the confidence threshold
## Safety guarantees
- **`engine = "none"` (default)**: Only extracts from parsed YAML frontmatter. No LLM involved, zero hallucination risk.
- **Confidence thresholds**: Facts below `min_confidence` are counted as "skipped" but never written.
- **Provenance**: Every fact records which meeting it came from and when.
- **Deduplication**: Facts whose text already appears in a person's profile are skipped.
- **Dry-run**: Always suggest `--dry-run` first if the user hasn't used ingest before.
## Interpreting the output
```
Ingesting 73 meeting(s) into knowledge base at /path/to/kb
2026-04-03-strategy.md — 4 written, 1 skipped — Mat, Dan
2026-04-05-standup.md — 2 written, 0 skipped — Alice
SKIP 2026-03-18-test.md: no frontmatter
Done. 6 fact(s) written, 1 skipped, 1 error(s), 3 people updated.
```
- **written**: facts that passed confidence threshold and didn't already exist
- **skipped**: facts below confidence threshold (logged, not written)
- **SKIP**: files that couldn't be parsed (no frontmatter, invalid YAML, etc.)
## Gotchas
- **Meetings without summarization have no structured data** — If a meeting was recorded before summarization was enabled, its frontmatter won't have `action_items` or `decisions`. The ingest will correctly extract 0 facts. This is expected, not an error.
- **`engine = "agent"` requires an AI CLI** — If the user wants richer LLM-based extraction from transcript body text, they need `claude`, `codex`, `gemini`, `opencode`, or `pi` on PATH.
- **PARA adapter writes `items.json`** — If the user's knowledge base uses the PARA format, facts go into `areas/people/{slug}/items.json` with atomic fact schema (id, status, supersededBy).
- **First run should be dry-run** — Always suggest `minutes ingest --all --dry-run` before the first real run so the user can see what would be extracted.Fast non-interactive briefing before any meeting — auto-detects your next calendar event, pulls relationship history, surfaces open commitments, and produces a one-page brief in under 30 seconds. Use this whenever the user says "brief me", "give me a quick brief", "what's coming up", "background on my next call", "who am I meeting next", "brief me on Sarah", "I have a call in 10 min", "quick rundown", or right before walking into a meeting. Different from /minutes-prep — brief is the fast hook-fireable version that doesn't ask questions and doesn't set goals. Use brief when speed matters; use prep when the user wants to think hard about goals first.
Manage old recordings — find large files, archive old meetings, delete processed originals. Use when the user says "clean up recordings", "how much space are meetings using", "delete old recordings", "archive meetings", "manage meeting storage", or asks about disk space from minutes.
Post-meeting debrief — analyzes what happened, compares outcomes to your prep intentions, tracks decision evolution. Use when the user says "debrief", "what just happened in that meeting", "what did we decide", "debrief that call", "post-meeting", "what changed", or right after stopping a recording.
Cross-meeting entity graph — query who/what/when across all your meetings as structured data, with co-occurrence and cross-entity queries that text search can't answer. Use whenever the user says "show me everyone who mentioned X", "all mentions of Y across meetings", "who knows about Z", "graph", "across all meetings", "entity search", "first time we talked about", "trend for X over time", "who's been mentioned alongside", or wants to query meetings as an index rather than full-text search. Builds a JSON entity index on first run (one-time slow), then answers queries instantly. Surface this skill for relationship intelligence, due diligence, or any "across all my history" question that text search alone can't answer.
Surface recent voice memos and ideas captured from any device. Use when the user asks "what ideas did I have?", "what were my recent memos?", "what did I record while walking?", or wants to recall a captured thought.
Health-check your meeting knowledge for contradictions, stale commitments, and decision conflicts. Use when the user asks "any conflicts in my meetings", "check for stale action items", "lint my meetings", "consistency check", "are there contradictions", or wants to audit their decision history.
List recent meetings and voice memos. Use when the user asks "what meetings did I have", "show my recent recordings", "any meetings today", "list my voice memos", or wants an overview of their meeting history. Also use when they need to find a specific meeting by browsing rather than searching.
Self-coaching analysis of your own behavior across meetings — talk-time ratio, filler words, hedging language, monologue length, energy patterns, and (when meetings are tagged via /minutes-tag) what your behavior in winning meetings looks like vs losing ones. Use this whenever the user says "how did I do", "review my last meeting", "mirror", "self-review", "show my patterns", "coach me", "where am I weak", "talk time", "am I improving", "what do I do in meetings I win", "feedback on me", or asks for any kind of personal feedback on their own meeting behavior. This is the rare skill that gives the user a mirror to their own habits — surface it whenever they show curiosity about their own performance, even if they don't use the word "mirror".