Canonical: https://botbento.com/usecases/keep-per-client-decision-notes-ready-for/
Format: Markdown representation of the public HTML page.

# Keep per-client decision notes ready for agent sessions

[Operations](https://botbento.com/usecases/ops/) · Intermediate

Resume client work without re-explaining past choices

Maintain a separate decision and context log for each client or project so an agent can pick up a new session, recall agreed rules, and avoid repeating rejected approaches. The bot reads recent project notes, updates a running brief, and surfaces open questions before work resumes.

## Quick answers

**What does it do?** Resume client work without re-explaining past choices. Result: Decision brief ready for review.

**What does it need?** Tools: GitHub, Docs, Notes. Intermediate: some setup. The prompt includes an approval step.

**How do I run it?** When something happens: New client session starts. Copy the prompt below into a bot in BotBento.

## How it works

1. When something happens: New client session starts
2. Then: Draft per-client context brief
3. Result: Decision brief ready for review

## Prompt to start with

```text
For each client or project, keep a separate running note of agreed rules, past decisions, rejected approaches, and open questions. When a session begins, read the latest notes, summarize what matters for the current task, and flag conflicts or missing context. Draft updates to the notes and a short brief for the user. Do not send messages, make payments, or delete notes without the user's approval.
```

## Tools it uses

Docs, Notes, GitHub

## Works with

BotBento

## Tags

context, client-work, decision-log, agents

## Source

[Ask HN: How do you manage your AI agents' context across clients and projects?](https://news.ycombinator.com/item?id=50403153) — Hacker News · checked 2026-11-09

Summarized in our own words. Any results described there are the authors’ own, not measured by BotBento.

Verified 2026-11-09: passed every automated gate and a second-model review (kimi-code/k3.1-pro). Automated, not human review. [How verification works](https://botbento.com/usecases/verification/)

By BotBento Editorial · Added 2026-11-09 · Updated 2026-11-09 · [Suggest a correction](https://botbento.com/contact/)

## Related use cases

- [Multi-agent software task coordinator](https://botbento.com/usecases/multi-agent-software-task-coordinator/) — Hand work between coding agents through one shared workspace instead of copy-paste
- [Resuming coding agent context from past sessions](https://botbento.com/usecases/resuming-coding-agent-context-from-past-sessions/) — Search historical coding logs to recover pre-compaction context and resume work
- [Filtering verbose terminal outputs for agent tasks](https://botbento.com/usecases/filtering-verbose-terminal-outputs-for-agent-tasks/) — Strip unnecessary logs and boilerplate from shell commands to keep context clear
