apex
This server does not currently have a deployable release on Glama. Follow the repository's setup instructions to install and configure it for your MCP client, including any required credentials and local dependencies.
APEX
Authorized security automation for modern web and AI systems.
APEX is a scope-first security automation platform that combines reconnaissance, web checks, secrets analysis, mobile analysis, LLM red-team testing, evidence collection and reporting behind explicit authorization gates.
Authorized use only. APEX is designed for bug-bounty programs, VDPs and engagements where the operator has explicit permission to test the target.
Why it matters
Security automation becomes dangerous when scope and authorization are treated as an afterthought. APEX makes them part of the execution model:
explicit scope definition
fail-closed target validation
authorization confirmation before active testing
per-host rate limiting
reproducible findings and evidence
Markdown/HTML reporting
MCP interface for AI-assisted workflows
The result is not just a collection of scanners, but an orchestration layer around specialized security agents and tools.
Related MCP server: keel
Architecture
┌─────────────────────┐
│ Scope / Policy │
│ fail-closed │
└──────────┬──────────┘
│
┌─────────▼─────────┐
│ Orchestrator │
└─────────┬─────────┘
│
┌────────────────────┼────────────────────┐
▼ ▼ ▼
Recon Web / API AI Security
│ │ │
Assets Findings agentstrike
└────────────────────┼────────────────────┘
▼
┌─────────────────────┐
│ Evidence / Quality │
│ / CVSS │
└──────────┬──────────┘
▼
Markdown / HTML / MCPCore capabilities
Area | Capability |
Scope | JSON programs, in/out-of-scope rules, authorization gates |
Recon | DNS/HTTP discovery and asset inventory |
Web | Security headers, TLS, exposed-file checks, CORS misconfiguration |
Takeover | Subdomain takeover: dangling-service fingerprint confirmed by CNAME to the service domain (recon) |
Open redirect | Parameter-based open-redirect check (non-destructive, no redirect followed) |
Secrets | Detection of exposed credentials with masked reporting |
Mobile | Offline APK analysis |
AI security | Prompt-injection testing through agentstrike |
Web vulnerabilities | Authorized SQLi/XSS checks through the web-vuln scanner |
Orchestration | Agent registry, dependencies, execution history |
Evidence | Quality gates, reproducible findings, CVSS 3.1 |
Reporting | Markdown and HTML reports |
AI integration | MCP server for controlled agent workflows |
Performance | Optional concurrent Go core with JSONL boundary |
Install
# recommended: isolated CLI install
pipx install apex-bounty
# or into the current environment
pip install apex-bounty
# with the optional MCP bridge
pip install "apex-bounty[mcp]"This installs the apex and apex-auto commands. The core has zero dependencies
(pure stdlib); [mcp] and [browser] extras are optional.
From source (development)
git clone https://github.com/nadirzhon/apex
cd apex
python3 -m apex.cli --help
pip install -e '.[dev]' # editable install + pytest/ruffQuick start
Create an explicit scope:
{
"program": "Example Corp",
"authorized": true,
"rate_limit_rps": 2,
"in_scope": ["*.example.com", "api.example.com"],
"out_of_scope": ["staging.example.com"]
}Run a baseline workflow:
apex --scope program.json scope
apex --scope program.json --i-am-authorized orchestrate --profile baseline --json
apex --scope program.json reportWithout authorization or when a target is outside scope, active execution fails closed.
New here? Follow the reproducible, offline demo walkthrough — it needs no network access or credentials.
MCP interface
APEX exposes controlled operations through MCP so an AI assistant can work with the same authorization boundary:
APEX_SCOPE=program.json APEX_AUTHORIZED=1 python -m apex.mcp_serverThe MCP layer can inspect scope, run authorized modules, review findings and generate reports without bypassing the policy gate.
Engineering highlights
Python standard-library core with optional integrations
Go concurrent network core for performance-sensitive workloads
deterministic scope enforcement
isolated modules and agent specifications
JSON state and JSONL event contracts
CVSS 3.1 scoring
evidence-quality validation
CLI + MCP interfaces
designed for reproducible security research rather than blind scanning
Related projects
agentstrike — LLM-agent red-team fuzzer
mcpscan — MCP security scanner
offsec-mcp — authorized security tools exposed to AI agents
License
MIT. The license does not remove the operator's responsibility to test only systems they are authorized to test.
Tool Schema Changelog
Recent tool additions, removals, and schema changes observed during successful MCP inspections.
Loading tool schema changelog…
This server cannot be deployed
Maintenance
Related MCP Connectors
Autonomous AppSec & DAST scanner across 12 security domains (260+ controls) with 3-Layer Triage.
AI pentesting: run scans, triage vulnerabilities, review PRs, manage schedules and assets.
Paid remote MCP for LLM security scans, jailbreak checks, analytics, checkout, and readiness.
Zero-config MCP security scanner for AI-generated apps. 25K+ vulnerability patterns.
Related MCP Servers
- AlicenseNot gradedqualityDmaintenanceEnables automated bug bounty hunting and security research with tools for reconnaissance, web vulnerability scanning, API testing, binary analysis, and mobile app analysis through an MCP interface.MIT
- AlicenseCqualityCmaintenanceEnables authorized pentest and bug bounty workflows from any MCP client, with scoped recon, per-host rate limits, and scanner output turned into deduplicated, triaged finding cards.919 PyPI2MIT
- AlicenseBqualityBmaintenanceEnables AI agents to conduct authorized pentests and bug-bounty workflows with scope enforcement, audit logging, reconnaissance, safe vulnerability checks, payload/wordlist generation, and professional report generation.14MIT
- AlicenseNot gradedqualityBmaintenanceEnables MCP clients to carry out a compliant, non-destructive, auditable authorized security-testing workflow — project intake, scope reading, asset discovery, non-destructive verification, evidence collection, and report-draft generation with manual submission tracking — where every outbound action is gated by scope checks, one-time human confirmation tokens, rate limits, and append-only audit logging.MIT