Nexus (OpenRouter Search)

Provides web search capabilities through OpenRouter's Perplexity Sonar models, delivering intelligent search results with source citations for current information retrieval.

ai-mlTypeScript
0Tools
10Findings
19Stars
Mar 19, 2026Last Scanned
5 critical · 3 high · 1 medium · 1 low findings detected

Security Category Deep Dive

Prompt Injection
Prompt & context manipulation attacks
69
Maturity
14
Rules
5
Sub-Categories
1
Gaps
64%
Implemented
56
Tests
1
Stories
PI-DIRDirect Input Injection
100%3 rules
Injection via tool descriptions and parameter fields
GAP-001Prompt Injection Coverage GapMissing detection coverage for emerging prompt injection attack variants not addressed by current rules
PI-INDIndirect / Gateway Injection
100%4 rules
Hidden instructions via external content and tool responses
PI-CTXContext Manipulation
100%2 rules
Context window saturation and prior-approval exploitation
PI-ENCEncoding & Obfuscation
100%3 rules
Payload hiding via invisible chars, base64, schema fields
PI-TPLTemplate & Output Poisoning
100%2 rules
Injection via prompt templates and runtime tool output
Framework Coverage
OWASP MCP Top 1014/14
MITRE ATLAS14/14
CoSAI MCP2/14
OWASP Agentic Top 1012/14
Kill Chain Phases
0Initial Access
0Defense Evasion
0Execution
0Persistence

Findings10

5critical
3high
1medium
1low

Critical5

criticalQ13MCP Bridge Package Supply Chain AttackMCP10-supply-chainAML.T0054
Pattern "["']@modelcontextprotocol/sdk["']\s*:\s*["'](?:\^|~|\*|latest)" matched in source_code: ""@modelcontextprotocol/sdk": "^" (at position 23846)
MCP bridge packages (mcp-remote, mcp-proxy, @modelcontextprotocol/sdk, fastmcp) are high-value supply chain targets — CVE-2025-6514 (CVSS 9.6) in mcp-remote affected 437,000+ installs. Always pin exact versions (no ^ or ~ ranges). Use lockfiles (package-lock.json, pnpm-lock.yaml, uv.lock). Never run `npx mcp-remote` without version pinning. Verify package integrity with `npm audit` or `pip-audit` before deployment. Reference: CVE-2025-6514, OWASP ASI04.
criticalK8Cross-Boundary Credential SharingMCP05-privilege-escalationAML.T0054
Pattern "(return|respond|output|result).*(?:token|credential|api[_\s-]?key|secret|password|bearer)" matched in source_code: "result.metadata.usage?.total_token" (at position 10120)
Never forward, share, or embed credentials across trust boundaries. Use OAuth token exchange (RFC 8693) to create scoped, delegated tokens instead of passing original credentials. Never include credentials in tool responses. Required by ISO 27001 A.5.17 and OWASP ASI03.
criticalL12Build Artifact TamperingMCP10-supply-chainAML.T0017
Pattern "(?:appendFileSync|appendFile|fs\.writeFileSync|>>).*(?:dist/|build/|out/|lib/).*\.(?:js|mjs|cjs)" matched in source_code: ">> dist/cli.tmp && mv dist/cli.tmp dist/cli.js && chmod +x dist/cli.js" (at position 19334)
Use SLSA provenance attestation to verify build artifact integrity. Publish with 'npm publish --provenance' for Sigstore attestation. Never modify files in dist/ after the build step. Use reproducible builds so artifacts can be independently verified. Separate build and publish into different CI jobs with artifact integrity checks. Enable npm audit signatures.
criticalQ11Code Suggestion Poisoning via MCPMCP01-prompt-injectionAML.T0054.001
Pattern "(?:suggest|generate|complete|insert).*(?:code|function|class|import|require)" matched in source_code: "Complete production build pipeline' && echo ' build:clean - Remove dist directory and artifacts' && echo ' build:compile - TypeScript compilation (no bundling)' && echo ' build:validate - Verify build output integrity' && echo ' build:cli-setup - Configure CLI executable' && echo '' && echo '🔧 DEVELOPMENT SCRIPTS:' && echo ' dev - Parallel development with watch & types' && echo ' dev:watch - Watch mode development server' && echo ' dev:types - TypeScript watch mode type checking' && echo ' dev:clean - Clean development artifacts' && echo '' && echo '🧪 TESTING SCRIPTS:' && echo ' test - Complete test suite (unit + integration)' && echo ' test:unit - Unit tests only' && echo ' test:integration - Integration tests only' && echo ' test:watch - Interactive test watch mode' && echo ' test:coverage - Tests with coverage reporting' && echo ' test:npx:local - Test local NPX installation' && echo ' test:npx:published - Test published NPX functionality' && echo '' && echo '🚀 RELEASE SCRIPTS:' && echo ' release - Complete release workflow' && echo ' release:patch - Automated patch release' && echo ' release:minor - Automated minor release' && echo ' release:major - Automated major release' && echo ' release:validate - Pre-release validation' && echo ' release:dry-run - Test publish without publishing' && echo '' && echo '✨ QUALITY SCRIPTS:' && echo ' lint - ESLint code quality checks' && echo ' lint:fix - Auto-fix linting issues' && echo ' format - Prettier code" (at position 21412)
MCP tool outputs flowing into IDE code suggestion contexts must be sanitized. Implement output content policies that: (1) strip hidden Unicode characters (zero-width, RTL override, tag characters), (2) detect embedded instructions targeting AI code assistants, (3) validate code blocks against security patterns before they enter the suggestion pipeline, (4) never include shell commands in tool outputs without explicit [COMMAND] markers visible to the user. Reference: IDEsaster (Dec 2025), arXiv 2509.22040.
criticalC1Command InjectionMCP03-command-injectionAML.T0054
Pattern "`[^`]+`" matched in source_code: "`AI-powered web search and knowledge retrieval via Perplexity Sonar and Grok models. MODEL SELECTION: - sonar: Fast Q&A, current web info (30s timeout) - sonar-pro: Multi-step research queries (60s timeout, premium) - sonar-reasoning-pro: Complex reasoning with citations (120s timeout, premium) - sonar-deep-research: Exhaustive multi-source reports (300s timeout, premium) - grok-4: Training data knowledge, no web search (60s timeout, premium) CONSTRAINTS: - Queries: 1-2000 chars - Response caching enabled (reduces API costs for repeat queries) - Concurrent identical requests automatically deduplicated - Premium models (sonar-pro+, grok-4) have higher API costs RESOURCES: - config://status: Server health and configuration info`" (at position 3111)
Replace exec()/execSync() with execFile() and pass arguments as an array, never as a string. Validate all inputs against an allowlist before use in any shell context. For subprocess.run, always pass a list and shell=False.

High3

highN10Incomplete Handshake Denial of ServiceMCP07-insecure-configAML.T0054
Pattern "(?:createServer|listen)\s*\((?!.*(?:maxConnections|maxClients|connectionLimit|MAX_CONN))" matched in source_code: "createServer(" (at position 16996)
Enforce a handshake timeout (recommended: 30 seconds) — terminate connections that do not complete the initialize handshake within the deadline. Limit maximum concurrent pending connections. An attacker can exhaust server connection slots by initiating MCP connections without completing the handshake (Slowloris-style attack). Reference: MCP spec 2025-03-26 lifecycle — initialize MUST complete before functional requests.
highK16Unbounded Recursion / Missing Depth LimitsMCP07-insecure-configAML.T0054
Pattern "(invoke|call|execute)[_\s-]?(?:tool|agent|self)(?!.*(?:depth|level|limit|max[_\s-]?(?:depth|recursi|iter|call)|count))" matched in source_code: "CallTool" (at position 192)
Add explicit depth/recursion limits to all recursive operations. Use iterative approaches where possible. Set maximum depth for directory walking (max_depth=10), tree traversal (max_level=20), and agent re-invocation (max_calls=5). Implement circuit breakers that halt after N iterations. Required by EU AI Act Art. 15 (robustness) and OWASP ASI08.
highK4Missing Human Confirmation for Destructive OpsMCP06-excessive-permissionsAML.T0054
Pattern "(delete|remove|drop|truncate|destroy|purge|wipe|erase).*(?:execute|run|perform|call)(?!.*(?:confirm|approve|prompt|ask|verify|consent))" matched in source_code: "Remove dist directory and artifacts' && echo ' build:compile - TypeScript compilation (no bundling)' && echo ' build:validate - Verify build output integrity' && echo ' build:cli-setup - Configure CLI executable' && echo '' && echo '🔧 DEVELOPMENT SCRIPTS:' && echo ' dev - Parallel development with watch & types' && echo ' dev:watch - Watch mode development server' && echo ' dev:types - TypeScript watch mode type checking' && echo ' dev:clean - Clean development artifacts' && echo '' && echo '🧪 TESTING SCRIPTS:' && echo ' test - Complete test suite (unit + integration)' && echo ' test:unit - Unit tests only' && echo ' test:integration - Integration tests only' && echo ' test:watch - Interactive test watch mode' && echo ' test:coverage - Tests with coverage reporting' && echo ' test:npx:local - Test local NPX installation' && echo ' test:npx:published - Test published NPX functionality' && echo '' && echo '🚀 RELEASE SCRIPTS:' && echo ' release - Complete release workflow' && echo ' release:patch - Automated patch release' && echo ' release:minor - Automated minor release' && echo ' release:major - Automated major release' && echo ' release:validate - Pre-release validation' && echo ' release:dry-run" (at position 21474)
All destructive operations (delete, drop, overwrite, send) MUST include a human confirmation step. Use the MCP destructiveHint annotation to signal that client-side confirmation is required. Implement an approval gate pattern: preview changes → request confirmation → execute. Required by ISO 42001 A.9.1, EU AI Act Art. 14, and NIST AI RMF GOVERN 1.7.

Medium1

mediumK20Insufficient Audit Context in LoggingMCP09-logging-monitoringAML.T0054
Pattern "logger\.(info|warn|error)\s*\(.*(?:tool|request|handle|invoke)(?!.*(?:requestId|correlationId|traceId|spanId|agent[_\s-]?id|user[_\s-]?id))" matched in source_code: "logger.info('Search tool" (at position 4133)
Use structured logging that includes all five ISO 27001 A.8.15 fields: (1) WHO — agent/user identity, (2) WHAT — tool name and operation, (3) WHEN — ISO 8601 timestamp, (4) WHERE — server ID and correlation ID, (5) OUTCOME — success/failure and result summary. Replace console.log with structured loggers (pino, winston). Add correlation IDs for request tracing across multi-agent chains.

Low1

lowF4MCP Spec Non-ComplianceMCP07-insecure-config
Server fails MCP spec compliance checks: required:server_name; required:server_version; required:protocol_version; recommended:tool_descriptions; recommended:parameter_descriptions
Follow the MCP specification for server metadata. Include server name, version, and protocol version. Provide descriptions for all tools and parameters.