Threat modeling and AI-reasoning vulnerability detection harness for Claude Code — STRIDE + AI + MAESTRO
This repository also configures its own agents. See what tachi tells them →
Cursor rule Cursor needs its repo
AI agent autonomy threat agent — detects excessive autonomy, goal misalignment, unconstrained action scope, missing human-in-the-loop checkpoints, cascading multi-agent failures, and autonomous resource consumption against Processes.
Cursor rule Cursor needs its repo
AI/LLM data poisoning threat agent — detects training data manipulation, RAG index poisoning, knowledge base corruption, fine-tuning supply chain attacks, and context window contamination against Data Stores and Data Flows.
Cursor rule Cursor needs its repo
STRIDE denial of service threat agent — detects availability degradation threats including resource exhaustion, algorithmic complexity attacks, connection pool exhaustion, cascading dependency failures, and application-layer flooding against Processes, Data Stores, and Data Flows.
Cursor rule Cursor needs its repo
STRIDE information disclosure threat agent — detects confidentiality violations including error message exposure, excessive data in responses, data at rest and in transit exposure, side-channel leakage, and debug endpoint exposure against Processes, Data Stores, and Data Flows.
Cursor rule Cursor needs its repo
AI/LLM model theft threat agent — detects model weight exfiltration, API-based model extraction, artifact exposure, side-channel reconstruction, fine-tuned model theft, and supply chain compromise against Data Stores and Processes.
Cursor rule Cursor needs its repo
Central coordinator for tachi OWASP threat modeling — parses architecture input, coordinates STRIDE and AI threat analysis, produces threats.md, SARIF output, and narrative threat report.
Cursor rule Cursor needs its repo
STRIDE elevation of privilege threat agent — detects unauthorized privilege gain including broken access control, insecure direct object references, role escalation, multi-tenancy boundary violations, and lateral movement against Processes.
Cursor rule Cursor needs its repo
AI/LLM prompt injection threat agent — detects direct and indirect prompt injection, jailbreak, system prompt extraction, and cross-plugin injection threats against LLM-integrated Processes.
Cursor rule Cursor needs its repo
STRIDE repudiation threat agent — detects accountability failures including missing audit trails, insufficient log detail, log tampering vulnerabilities, timestamp manipulation, and deniable actions against External Entities and Processes.
Cursor rule Cursor needs its repo
STRIDE spoofing threat agent — detects identity impersonation, authentication bypass, credential theft, session hijacking, and federated identity attacks against External Entities and Processes.
Cursor rule Cursor needs its repo
STRIDE tampering threat agent — detects unauthorized data modification threats including input injection, data flow manipulation, persistent data corruption, code and configuration tampering, and supply chain attacks against Processes, Data Stores, and Data Flows.
Cursor rule Cursor needs its repo
Threat infographic generator — transforms structured threat model output into visual infographic specifications and images via Gemini API. Supports multiple templates: Baseball Card (risk summary dashboard) and System Architecture (annotated architecture diagram with attack surface badges).
Cursor rule Cursor needs its repo
Threat report generator — produces narrative threat analysis with executive summary, Mermaid attack trees for Critical and High findings, prioritized remediation roadmap with effort estimates, cross-cutting theme detection, and complete finding traceability.
Cursor rule Cursor needs its repo
AI tool-use abuse threat agent — detects unauthorized tool invocation, capability escalation through tool composition, parameter injection, tool chain manipulation, and tool poisoning attacks against Processes with MCP servers, plugins, or function calling.
At most 3 mods per repository are shown here, and a mod shipped inside a plugin is left to that plugin's page — the rest are on their repository pages: