Trust score
30/100
MarketNow Registry / Developer Tools / agent-threat-rules
agent-threat-rules
Open detection standard -- like Sigma, but for AI agents. Executable rules for prompt injection, tool poisoning, context exfiltration, and MCP attacks. Shipped in Cisco AI Defense. MIT-licensed.
Developer Tools
v4.0.0
by panguard0414
Install risk
Runs registry code
Downloads / week
7,615
Source
npm-registry
Indexed
2026-09-10
Install
$ npx -y agent-threat-rulesInstalling via npx/uvx/npm install executes arbitrary code from the public registry on your machine. This is the install-risk semantic of the MarketNow catalog — verify the publisher before running it.
Adoption evidence: 7,615 weekly registry downloads.
How MarketNow scored this
Sentinel Index Heuristics — a security-first estimate from public signals:
- Package age and release activity (stale repositories lose points)
- Verified adoption: registry downloads or GitHub stars, not follower counts
- Typosquat distance against the 150 most-installed MCP packages
- Injection markers scanned in the package description
Heuristic, not a guarantee. For verified agent credentials and revocation, see the MarketNow Trust API.
What is agent-threat-rules?
Open detection standard -- like Sigma, but for AI agents. Executable rules for prompt injection, tool poisoning, context exfiltration, and MCP attacks. Shipped in Cisco AI Defense. MIT-licensed. Indexed by MarketNow with trust 30/100 (Runs registry code).
How do I install it?
Run $ npx -y agent-threat-rules Installing via npx/uvx/npm install executes arbitrary code from the public registry on your machine. This is the install-risk semantic of the MarketNow catalog — verify the publisher before running it.
What does the trust score mean?
Security-first heuristic over public signals (age, adoption, typosquat distance, injection markers). It is not a popularity ranking and not a guarantee of safety. agent-threat-rules scores 30/100.