Install, run the demo, capture one session, then query one decision with cited local evidence.
Roll work up for reporting. Drill down to verify.
Lode turns scattered work from AI collaboration into report-ready, verifiable outcomes. When questions arise, trace each outcome back to decisions, tradeoffs, and evidence.
Report upward: session signals → work streams → outcomes and impact
Verify downward: outcomes and impact → decisions and tradeoffs → source refsThis use case is for developers, tech leads, and small technical teams that need to explain technical work. Activity counts measure record coverage; they do not prove outcomes by themselves. Lode does not expand into generic meetings, approvals, or employee monitoring.
Each skill maps to a habit you already have. Activate it with a namespaced command.
| Skill | When | What it does |
|---|---|---|
/lode:cold-start-interview | First run | Creates ~/.lode/config.yaml with vault path, project identity, and report preferences |
/lode:capture | Wrap-up or checkpoint | Quietly captures session depth, stage progress, and artifacts |
/lode:query | Targeted follow-up | Answers "why did we choose this?" with cited decision replay evidence |
/lode:recall | Session start, after history exists | Recalls recent decisions, risks, open questions, and relevant docs |
/lode:roadmap | After multiple decisions | Generates narrative decision roadmap with accumulating risks |
/lode:daily | Daily, on demand | Updates Obsidian daily notes from raw entries and git history |
/lode:weekly | Weekly, on demand | Builds weekly outline from raw entries with conditional hard-stuff section |
/lode:monthly | Monthly, on demand | Generates monthly review and candidate rules from repeated evidence |
install → demo → capture one session → query one decisionThis is the first value path. Weekly, monthly, recall, and roadmap outputs are compounding layers on top of the same local record. Skills are independent, but they share one storage convention so downstream views reuse earlier evidence.
Lode dogfoods a derived decision replay index for its own project history. Raw entries remain the source of truth; {vault}/raw/decisions/{slug}.json gives coding agents compact evidence packs for targeted "why did we choose this?" queries.
See the dogfood proof on GitHub.
codex plugin marketplace add KKenny0/Lode
codex plugin add lode@lodeUpdate:
codex plugin marketplace upgrade lode
codex plugin add lode@lodeOnly use the legacy fallback when codex plugin add is unavailable:
npx @lode/cli install-codex-pluginThen run the demo, run /lode:cold-start-interview once, capture one real session with 收工, and query one decision with /lode:query.
No vault? No problem. 收工 outputs structured Markdown directly in the conversation. With a vault, capture writes quietly by default; use /lode:capture checkpoint during long sessions.