Agent Security Review

advanced

Audit agent tools and skills for prompt injection, secret exposure, and unsafe execution paths.

Architecture

Scans agent skills and MCP surfaces for attack classes

78/100

npx -y arc-security-mcp

Static Analysis Skill

Finds unsafe code patterns in agent implementation

90/100

ls /workspace

Secrets Detection Skill

Checks configs and examples for leaked credentials

85/100

ls /workspace

Threat Modeling Skill

Turns risks into prioritized STRIDE-style mitigations

80/100

ls /workspace

Expected Output

Agent Security Review Tool surfaces: 11 Prompt-injection risks: 4 Secret leaks: 0 Critical mitigations: 2 -> Review packet ready for launch gate

Related Recipes

Full-Stack Testing Suite

Complete testing pipeline with unit tests, E2E browser tests, and TDD workflow.

3 components

API Development Workflow

Complete API lifecycle from design to testing with GraphQL support.

3 components

Visual Technical Briefing

Create architecture diagrams, editable whiteboards, and browser-ready briefing decks.

4 components