Agent touchpoints

Give your team one agent they can reach from the tools they already use.

Tag it in Slack, open or update work in GitHub, fire an incident, or emit an internal event. ExecWarden carries that context into one governed workflow and keeps each response inside a separate grant.

One agent, many touchpoints

It can feel present in every system without becoming unrestricted in any of them.

Slack messages and mentions can enter through the configured Slack event receiver; GitHub events use the supported GitHub path; incident tools, CRMs, and internal applications can use signed webhook receivers or declared package events. They can all reach the same governed workflow and Agent identity.

Slack

@mention or channel message

Configured Slack Events API receiver

Read or post in allowlisted channels; posting can require Ask approval.

GitHub

Pull request, issue, or push

Supported GitHub event

Review, comment, or open a PR through a repository-scoped GitHub grant.

Incident tools

Alert or incident callback

Signed webhook receiver

Investigate with granted context; update or deploy only through separate capabilities.

CRM or internal system

Record or domain event

Signed webhook or declared package event

Read or write through the selected package or API action, with approval where required.

1

The touchpoint sends context

A mention, pull request, alert, or record change becomes a normalized event.

2

The same named Agent responds

One workflow and Agent identity carry the job across every entry point.

3

Each answer has its own grant

Posting, commenting, updating, or deploying stays separately scoped and reviewable.

Run evidence

Follow an incoming event into the resulting work.

Delivery status and workflow-run status stay visible together, including work that completes, pauses for a person, or fails before a run can start.

Workflow run history with waiting, succeeded, and failed runs beside started and failed incident event deliveries
The incident example links received events to workflow runs while keeping waiting and failed outcomes visible.

Examples

What one agent presence can look like across your systems.

1A teammate asks in Slack why a deployment failed. The agent receives the thread context, inspects granted GitHub and CI data, then asks before posting its answer.
2A pull request opens in GitHub. The agent reviews the diff, comments through its repository-scoped grant, and can send a Slack handoff only when that channel action is also granted.
3An incident callback starts an investigation. The agent can prepare a GitHub fix and an incident update, while each external write keeps its own approval decision.

Start where the work happens

Connect one touchpoint and let the same governed agent respond.

Start with GitHub, a configured Slack app, or a signed callback from an incident tool, CRM, or internal application.

Create an event workflowSee the technical path