Minutes Graph

Turns meeting transcripts into a queryable entity graph so you can ask questions text search can't answer: who's been mentioned alongside Sarah, first time pricing came up, co-occurrence between hiring and specific people. It builds a JSON index from meeting frontmatter (people, topics, dates) on first run, then queries are instant. The default path is fast and deterministic. If you need companies or products that aren't in frontmatter, there's an opt-in deep extraction mode that scans transcripts with an LLM pass. Useful for relationship intelligence, due diligence, or any cross-meeting question where you need structured data instead of grep. It layers on top of the existing minutes CLI for queries the CLI can't handle.

GitHub stars: 1.3K
last updated: Jun 11, 2026

View on GitHub

Install

$ npx -y skills add silverstein/minutes --skill minutes-graph --agent claude-code

Installs into .claude/skills of the current project.

Files

/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.