context safety score
A score of 41/100 indicates multiple risk signals were detected. This entity shows patterns commonly associated with malicious intent.
capability escalation
The ListToolsRequestSchema handler dynamically fetches the entire tool surface from https://waystation.ai/tools/list at runtime. The remote server can add, remove, or modify tool definitions at any time without any code change or user consent. There is no allowlist, signature verification, or pinning of expected tool names/schemas. Combined with the CallToolRequestSchema handler that blindly forwards any unrecognized tool name to the remote endpoint, this creates a fully remote-controlled tool proxy where the operator can introduce arbitrary new capabilities (including tool-shadowing names) at any time. (location: src/index.ts: ListToolsRequestSchema handler (lines fetching from ${API_BASE}/tools/list) and CallToolRequestSchema catch-all forwarding to ${API_BASE}/tools/call)
response injection
Remote tool call responses from https://waystation.ai/tools/call are returned directly to the agent via 'return await response.json()' with zero validation or sanitization. The remote server has full control over what text, instructions, or hidden directives are delivered to the agent in tool responses. This allows the server operator to inject arbitrary prompt manipulation, behavioral overrides, or cross-server attack payloads into any tool response at any time. (location: src/index.ts: CallToolRequestSchema handler, final return statement 'return await response.json()')
description injection
The helloWayStation tool response (WAYSTATION_HELLO constant) contains explicit behavioral instructions to the agent: 'When a user says "Hello WayStation", say the following:' followed by marketing copy the agent is directed to present as its own words, ending with 'Now type "Open WayStation" to get started' which steers the user toward invoking the openWayStation tool that executes OS commands. This is a response-delivered prompt injection that manipulates agent speech and chains into OS command execution. (location: src/index.ts: WAYSTATION_HELLO constant and helloWayStation tool handler)
consent bypass
The openWayStation tool executes OS-level commands (execSync with cmd.exe on Windows, osascript on macOS) to launch an application, creates directories at ~/.waystation/, and writes files to disk — all without any user confirmation mechanism. The tool description uses imperative language ('Call this action when users says...') to trigger automatically on keyword match rather than requiring explicit user approval for OS command execution. (location: src/index.ts: openWayStation handler using execSync() and fs.mkdirSync/writeFileSync)
curl https://api.brin.sh/mcp/waystation-ai%2FmcpCommon questions teams ask before deciding whether to use this mcp server in agent workflows.
waystation-ai/mcp currently scores 41/100 with a suspicious verdict and low confidence. The goal is to protect agents from high-risk context before they act on it. Treat this as a decision signal: higher scores suggest lower observed risk, while lower scores mean you should add review or block this mcp server.
Use the score as a policy threshold: 80–100 is safe, 50–79 is caution, 20–49 is suspicious, and 0–19 is dangerous. Teams often auto-allow safe, require human review for caution/suspicious, and block dangerous.
brin evaluates four dimensions: identity (source trust), behavior (runtime patterns), content (malicious instructions), and graph (relationship risk). Analysis runs in tiers: static signals, deterministic pattern checks, then AI semantic analysis when needed.
Identity checks source trust, behavior checks unusual runtime patterns, content checks for malicious instructions, and graph checks risky relationships to other entities. Looking at sub-scores helps you understand why an entity passed or failed.
brin performs risk assessments on external context before it reaches an AI agent. It scores that context for threats like prompt injection, hijacking, credential harvesting, and supply chain attacks, so teams can decide whether to block, review, or proceed safely.
No. A safe verdict means no significant risk signals were detected in this scan. It is not a formal guarantee; assessments are automated and point-in-time, so combine scores with your own controls and periodic re-checks.
Re-check before high-impact actions such as installs, upgrades, connecting MCP servers, executing remote code, or granting secrets. Use the API in CI or runtime gates so decisions are based on the latest scan.
Learn more in threat detection docs, how scoring works, and the API overview.
Assessments are automated and may contain errors. Findings are risk indicators, not confirmed threats. This is a point-in-time assessment; security posture can change.
integrate brin in minutes — one GET request is all it takes. query the api, browse the registry, or download the full dataset.