Two ways to discover an agent
Tool pages help you discover useful jobs for software you already use. Workflow pages show how several services, agents, and approval points connect to deliver one finished outcome.
AI agents by tool
Ideas, common service pairings, connector options, and safe operating boundaries.
Team communication
Research, triage, and prepare decisions before delivering the result to the right Slack channel.
6 jobs to explore
Software development
Review repository activity, investigate failures, and prepare safe, evidence-backed changes.
8 jobs to explore
Knowledge management
Turn workspace knowledge into maintained databases, research briefs, and reliable team follow-through.
7 jobs to explore
Triage inboxes, prepare evidence-backed drafts, and keep sending under explicit human control.
6 jobs to explore
Product development
Triage product work, surface stale decisions, and turn repository or customer evidence into cleaner issues.
8 jobs to explore
Documents
Turn changing folders and documents into structured briefs, checks, and maintained team knowledge.
6 jobs to explore
AI agent workflows
Detailed tutorials for connecting services, mapping agent responsibilities, preparing context, and generating a customized implementation.
Finance operations tutorial
Make the workflow explicit, prepare its context and data contract, then give a coding agent one prompt to create the right AgentUse agent—or small agent team—for your environment.
SEO operations tutorial
For SEO teams that want monitoring without preparing exports by hand, this tutorial collects a controlled weekly crawl and ranking snapshot from DataForSEO, compares it with the prior run, and prepares the changes that need review.
Engineering operations tutorial
For teams with a small set of business-critical libraries, this tutorial monitors only configured upstream repositories, compares new releases and advisories with internal usage, and prepares evidence-backed follow-up drafts.
The AgentUse pattern
Instructions, model, tools, and guardrails live in a readable Markdown file in your repo.
Give the agent narrowly scoped tools through MCP, the CLI, or your own integration.
Run from schedules, HTTP, CI, or Docker—with sessions and approvals in AgentUse Serve.