Canonical: https://botbento.com/usecases/internal-chat-agent-that-scopes-every-request/
Format: Markdown representation of the public HTML page.

# Internal chat agent that scopes every request to the person asking

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

Each request runs under the asker's own permissions, not a shared key

An internal chat agent that handles lookups and record changes for many people at once, treating each request according to who made it rather than through one shared secret everyone inherits. It keeps the requester's identity attached to everything it does, declines work outside that person's scope, and leaves a readable trail of what was done and on whose behalf.

## Quick answers

**What does it do?** Each request runs under the asker's own permissions, not a shared key. Result: Scoped result plus audit entry.

**What does it need?** Tools: Slack, Chat, Database. Intermediate: some setup. The prompt includes an approval step.

**How do I run it?** When you ask: Team member asks the bot for something. Copy the prompt below into a bot in BotBento.

## How it works

1. When you ask: Team member asks the bot for something
2. Then: Resolve identity and check scope
3. Result: Scoped result plus audit entry

## Prompt to start with

```text
When a team member asks you to look something up or change a record, first resolve who is asking and act only with that person's own permissions. Never fall back to a shared credential on someone else's behalf. If the request is outside the asker's scope, decline and say why. Draft any outgoing message, payment, or deletion and wait for explicit approval before doing it. Record each action, the identity behind it, and the outcome so it can be reviewed later.
```

## Tools it uses

Chat, Slack, Database

## Works with

BotBento

## Tags

access-control, multi-user, audit-log, chat-ops, policy

## Source

[Launch HN: Alter (YC S25) – Access control for multi-user agents](https://news.ycombinator.com/item?id=50176908) — Hacker News · checked 2026-11-13

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

Verified 2026-11-13: 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 2026-11-13 · Updated 2026-11-13 · [Suggest a correction](https://botbento.com/contact/)

## Related use cases

- [Triage agent access requests with scoped approvals](https://botbento.com/usecases/triage-agent-access-requests-with-scoped-approvals/) — Check purpose and scope before an agent gets tool access
- [Explain rule-based case decisions with cited policy passages](https://botbento.com/usecases/explain-rule-based-case-decisions-with-cited/) — Evaluate case facts against fixed rules, then draft a cited rationale
- [Uncovering user friction and feature asks in chat transcripts](https://botbento.com/usecases/uncovering-user-friction-and-feature-asks-in/) — Identify agent failure patterns, user corrections, and unhandled feature requests from chat logs.
