Mimir — Security agent for Claude Code
External research specialist - web, GitHub, documentation, best practices, security advisories.
How to install Mimir
Installs to ~/.claude/agents/lizardliang-kratos-mimir.md
mkdir -p ~/.claude/agents && curl -fsSL https://raw.githubusercontent.com/LizardLiang/kratos/HEAD/agents/mimir.md -o ~/.claude/agents/lizardliang-kratos-mimir.md Restart Claude Code, or start a new session, for it to be picked up.
What Mimir does
name: mimir description: External research specialist - web, GitHub, documentation, best practices, security advisories tools: WebFetch, WebSearch, Bash, Read, Write, Edit, Glob, Grep, Task model: sonnet model_eco: haiku model_power: opus protocol_sections: auto-discovery, missing-required-input, plain-language, boundaries, output-format
Mimir - God of Knowledge (Research Oracle)
You are **Mimir**, the all-knowing research oracle. You gather knowledge from the outside world - web, Gi
Alternatives in Security
- Gitnexus Security Boundary Reviewer — GitNexus security and trust-boundary reviewer 45.8k ★
- Malware Analyst — You are an elite malware analyst focused on defensive security research 31.9k ★
- Security Analyzer (Synthesizer) — You consolidate findings from research agents into prioritized hypotheses 215 ★
Full documentation available on GitHub
View Source RepositoryRelated Agents
Osf Researcher
Research specialist. Searches the web for technical information, best practices, documentation, comparisons, a
Cspec Research
Research agent for /cspec. Searches the web for current best practices, recent changes, security advisories, a
Sweep Lane Web
A scanning lane of this repository's security sweep whose brief needs published documentation or advisories, l
Negative Test Planner
Designs negative, validation, and security test cases (TC-NEG-NNN) from confirmed requirements, backed by web
Ext Attack Planner
Reasons from an external-pentest inventory to the most likely footholds in an authorized engagement. Correlate
Audit Doc Writer
Generates focused audit document content for a specific topic. Has web access for build-vs-buy research.
Related Skills
Harden Codebase Properly
A Claude Code skill: measure an existing web app, repair it, and keep it at a verifiable production bar — gate
AI Trust
VirusTotal for AI packages. ai-trust check returns a 0-100 trust score from the OpenA2A Registry trust graph,
Advisories
Triage and clear known security advisories by reachability — batched so a break is attributable, nothing bundl