Slack agent terms.

Scope & definitions.

These terms apply in every Slack channel where Stacklumen agents operate — our internal workspace and the shared channels we run with clients and partners.

// Agents

AI assistants operating inside Stacklumen’s Slack workspace and shared client channels. Agents can read requests, draft responses, and — under staff direction — perform site and platform operations through Stacklumen’s internal tooling.

// Internal tooling

The agency-controlled integrations agents use to make changes, including our website-platform integration. This tooling lives inside Stacklumen’s workspace and is available only to Stacklumen staff.

// Who this covers

Stacklumen staff, and every client, partner, or guest who participates in a Slack channel where agents operate.

// Other policies

These terms sit alongside our Legal & Policies and App Terms. Where they overlap, the stricter protection applies.

How a change request works.

You never talk to an agent. You talk to us — in plain language, in your shared channel — and we run the agentic work internally.

Access & authority.

Agentic tooling is internal by design. Restricting who can direct the agents is what makes agent-shipped changes safe enough to offer at all.

// Internal only

Agentic tooling — including integrations that can modify live sites — is restricted to Stacklumen staff. It is not available to clients, partners, or third parties, in any channel.

// Why

Two reasons: legal accountability for changes to production systems, and protection of our proprietary prompt and tooling systems.

// No commitments

Agent messages are operational assistance, not offers or contracts. Pricing, scope, and commitments are binding only when confirmed by Stacklumen staff.

// Revocation

We may pause, limit, or withdraw agent operations in any channel at any time, without notice.

Data & privacy.

Agent-operated channels are still your channels. Here’s what gets processed, what doesn’t belong in Slack, and how retention works.

// What’s processed

Messages in channels where agents operate may be processed by third-party AI providers to generate responses and perform requested work.

// What not to post

Never share passwords, API keys, full payment-card numbers, or regulated health information in Slack. If a change requires credentials, we’ll provide a secure handoff.

// Retention

Slack content follows your workspace’s retention settings. We don’t use your content to train models of our own; AI providers process it under their commercial terms.

// Alignment

Data handling follows our Privacy Policy. Where these terms are stricter, these terms apply in agent-operated channels.

Accuracy & responsibility.

Agents make us faster; they don’t replace judgment. Staff review sits between agent output and anything that reaches production.

// Assistive output

Agent responses can contain errors. Staff verify agent output before changes reach production.

// Published changes

Live-site changes ship only after staff review. If something ships wrong, flag it in the thread — we fix it.

// Not advice

Nothing an agent writes is legal, financial, or professional advice.

// Availability

Agents may be offline or limited at any time. The workflow falls back to humans — the @Stacklumen tag always works.

Acceptable use.

A few hard lines. They protect your site, your data, and every other client who shares the workspace with us.

// No direct use

Don’t attempt to direct, trigger, or converse with agents to perform work — including crafted messages designed to be picked up by an agent.

// No injection

Don’t attempt prompt injection, jailbreaks, or instructions hidden in messages, files, or links shared into agent-operated channels.

// No extraction

Don’t use agents to attempt access to Stacklumen’s internal instructions, tooling, other clients’ data, or any system you’re not authorized to use.

// Enforcement

Violations may lead to removal from shared channels and are covered by the enforcement terms in our App Terms.