checks: would an ATR-based agent-threat detection check be in scope for the Hub?

[eeee2345]

Hi maintainers,

Opening an issue first to gauge appetite before any code work.

Question

Would maintainers accept a contributed AI check that scans LLM input/output against ATR (Agent Threat Rules) and gates the result as a PR status check, in the same shape as the security review example in the README?

The intended contribution:

• One self-contained markdown check (the same format as .continue/checks/ examples)
• Optionally shareable via the Continue Hub so other repos can pull it in
• Wraps the ATR rule set (425 open MIT-licensed YAML rules) into the check's evaluation prompt and CI gate
• Does not modify Continue core; purely additive

Why this might be in scope

Agent-threat detection in PRs is a real gap for IDE-assistant deployments, and the README already shows security-flavored checks as a first-class use case. The same kind of rule pack is already running in production at Microsoft Agent Governance Toolkit (PRs #908 and #1277, both merged) and Cisco AI Defense skill-scanner (PRs #79 and #99, both merged), so the detection layer is field-tested. ATR is MIT-licensed throughout, so no IP friction for a check distributed through Continue.

What I would draft if you say yes

• The check markdown file (and a Hub entry if that surface is the right home)
• Integration docs explaining install + CI wiring
• A small test fixture covering the check against 5-10 representative attack samples

Repo for the rules: https://github.com/Agent-Threat-Rule/agent-threat-rules

If you say no

Totally fine. No PR will follow. I will not bump or repost.