Who it is forAn evaluation engineer who needs to compare Argus against other agents on the same Harbor tasks.
The problemComparing agent systems fairly requires running each one through the same harness, and Argus is not wired into Harbor by default.
The outcomeHarbor results that record how the bounded Argus runtime performed on the chosen tasks.
What you need first
- Node.js 22.12 or later
- One authenticated Agent CLI, such as GitHub Copilot CLI, OpenAI Codex CLI, or Claude Code
- A project folder you are willing to let the agents modify
Steps
- Install Docker, which the README lists as required only for the Harbor integration.
- Complete the basic Argus install with Node.js 22.12 or later and one authenticated Agent CLI.
- Open the Harbor integration guide at docs/harbor.md.
- Register Argus as a custom agent in Harbor by following that guide.
- Run your Harbor task set and read the results alongside the Argus status output.
Prompt for your agent
Using the microsoft/ArgusAgent repo, follow docs/harbor.md to run the Argus runtime as a custom agent in Harbor Framework on the task set <task set name>. Expect Harbor's result report for each task. Verify that the runtime ran in its bounded form and that each task result links back to its Argus run.
Caveat. The README describes Harbor as a separate integration, so it is not part of a normal Argus install and needs its own guide.
Use this in your AI agent
Ask your agent
Using the Deduped MCP, get the setup steps for microsoft/ArgusAgent and "Benchmark the Argus runtime as a custom agent in Harbor"
Add Deduped to Claude Code once
claude mcp add --transport http deduped https://dedupedai.com/mcp
Every Monday: the 5 hottest repos and exactly how to use them. Free, one email a week, unsubscribe anytime.
Get them freeShare: X LinkedIn Hacker News Email Copy link
Third-party repos we don't own. AI-written use cases can be wrong. Disclaimer