Interest Radar: skill pack for Hermes Agent

Platform-agnostic skill that judges how relevant a link or piece of content is to you

What Interest Radar does

Interest Radar is a relevance-judgment skill: given an external link or piece of content, it tells you how relevant it is to you, why, and whether it is worth reading. It does not collect, push or store anything. The workflow is defined entirely in SKILL.md, with no Python scripts or external runtime, and it is not tied to one platform; the README lists Hermes, Claude Code, Codex, Cursor and ChatGPT as places it can run.

At run time it detects whichever memory and search tools the agent has and uses them to recall context. It offers three capabilities: forward-and-judge, where you send a link and your intent; batch_judge, which other push skills can call to filter and deliver only high-relevance items; and a feedback loop where replying confirm or dismiss adjusts later scoring. Reference files cover the batch interface, the feedback loop, the output format and the snapshot cache. The README is written in Chinese.

Key features

Relevance judgment with an explanation of why and whether to read
Pure SKILL.md workflow with no scripts or external runtime
Adaptive context recall that detects available memory and search tools
batch_judge interface for other push skills to filter content
Confirm or dismiss feedback that adjusts later scoring
Reference docs for batch interface, feedback loop, output format and snapshot cache

When to use it

Forwarding a link to your agent and asking whether it is worth your time
Filtering a batch of feed items so only high-relevance ones get pushed
Training an agent's judgment over time by confirming or dismissing its calls

Who it is for: Hermes users who read a lot of links and want an agent to screen them for relevance, and can work from a Chinese-language README.

How it fits with Hermes Agent

A Hermes skill that installs by copying into ~/.hermes/skills/research/, and the README says it also runs on other agent platforms.

How to install Interest Radar

These commands are copied from the project's README. Check the repository for the latest steps before you run them.

cp -r interest-radar ~/.hermes/skills/research/interest-radar

Note: The README is written in Chinese, and the repository has no license file.

FAQ

What is Interest Radar?

Interest Radar is a general relevance-judgment engine defined as an agent skill. It does not collect, push or store content; it only tells you how relevant a piece of content is to you and why.

Does Interest Radar work with Hermes Agent?

Yes. The README gives a Hermes install path under ~/.hermes/skills/research/. It is also platform-agnostic and can run on Claude Code, Codex, Cursor and ChatGPT.

How do I install Interest Radar?

Copy SKILL.md and the references folder into your agent's skills directory. For Hermes, the README uses cp -r interest-radar ~/.hermes/skills/research/interest-radar.