Mythos Agent
by mythos-agentcommunityv4.0.1
14
2
mythos-agent is an AI-assisted security code review tool that scans source code for likely vulnerabilities, security misconfigurations, secrets, and known-CVE variants. It can run autonomous multi-stage hunts, verify findings with fuzzing/PoC generation, and suggest fixes or export reports.
From the registry: Open-source AI security agent: SAST, DAST, and policy-as-code over MCP.
Install & configure
No canonical install info was extracted for this server. Check the repository README directly.
Tools (13)
hunt
Runs a full autonomous multi-agent security scan through reconnaissance, hypothesis generation, analysis, and exploit/reproduction stages.
scan
Performs a standard security scan using patterns, secrets, dependency, IaC, and AI-based checks.
variants
Finds code variants structurally similar to a given CVE root cause.
fix
Generates AI patches for findings, optionally applying them.
ask
Answers natural-language security questions about the codebase.
taint
Performs AI-based data flow and taint analysis.
watch
Continuously scans on file save.
dashboard
Opens a local web UI with charts and a findings table.
report
Exports results as terminal output, JSON, HTML, or SARIF.
policy
Manages policy-as-code for compliance checks such as SOC2, HIPAA, PCI, and OWASP.
rules
Searches, installs, and publishes community rule packs.
tools
Checks which external security tools are installed.
init
Runs a setup wizard for connecting AI providers such as Anthropic, OpenAI, Ollama, or LM Studio.
Packages
mythos-agent
npmv4.0.1transport: stdio
Details
- Language
- TypeScript
- License
- MIT
- Released
- Apr 24, 2026
- Last commit
- Apr 27, 2026
- Registry name
- io.github.mythos-agent/mythos-agent