More screenshots
Works with the AI Agents you already use
LLM App Safety Check
exec(response.output_text), a NEXT_PUBLIC_OPENAI_API_KEY in .env.local, or f"SELECT ...
$5
LLM App Safety Check
Example session with this skill installed
Check this project with llm-app-safety-check.
- Read your context and instructions
- Compiled the llm app safety
$ python3 scripts/check_llm_app.py examples/risky
BLOCK .env.local:3 L04 KEY_EXPOSED_TO_BROWSER
NEXT_PUBLIC_GEMINI_API_KEY=<redacted>
Why: An LLM API key reaches the browser: a NEXT_PUBLIC_, VITE_, EXPO_PUBLIC_, REACT_APP_, NUXT_PUBLIC_, PUBLIC_, GATSBY_ or VUE_APP_ variable that holds a provider key (set in a .env file, read in code or passed as apiKey to an LLM client), or an OpenAI or Anthropic client made with dangerouslyAllowBrowser: true and a key from the build. Bundlers copy these values into the JavaScript every visitor downloads, so anyone can copy the key from the browser's developer tools.
Safer: Call the model from a server route (a route handler, server action or API) that reads a server-only variable such as OPENAI_API_KEY, rotate the exposed key, and have the browser call your route. For realtime voice in the browser, mint a short-lived client secret on the server.
BLOCK agent/report_agent.py:8 L01 API_KEY_IN_CODE
BLOCK agent/report_agent.py:20 L03 MODEL_OUTPUT_IN_SQL_TEXT
BLOCK agent/report_agent.py:27 L02 MODEL_OUTPUT_RUN_AS_CODE
BLOCK lib/gemini-browser.ts:4 L04 KEY_EXPOSED_TO_BROWSER
REVIEW agent/report_agent.py:13 L12 NO_OUTPUT_TOKEN_LIMIT
REVIEW agent/report_agent.py:14 L13 PROMPT_OR_RESPONSE_LOGGED
REVIEW agent/report_agent.py:25 L16 MODEL_WRITTEN_SQL_RUN
REVIEW agent/tools.py:9 L15 DANGEROUS_AGENT_TOOL
REVIEW app/api/chat/route.ts:7 L13 PROMPT_OR_RESPONSE_LOGGED
REVIEW app/api/chat/route.ts:9 L12 NO_OUTPUT_TOKEN_LIMIT
REVIEW app/api/chat/route.ts:12 L11 USER_INPUT_IN_SYSTEM_PROMPT
REVIEW components/ChatPanel.tsx:15 L14 MODEL_OUTPUT_AS_HTML
(Each finding also has Why and Safer lines, as in the first one.)
Verdict: BLOCK. Checked 6 files, 6 with LLM app content.
This is a static check of file text. NOT FLAGGED is not a guarantee.
Connects securely to your tools. The creator never sees your data.
About this skill
exec(response.output_text), a NEXT_PUBLIC_OPENAI_API_KEY in .env.local, or f"SELECT ... WHERE category = '{category}'" with a category the model chose: each one turns a single prompt injection, or one look at your JavaScript bundle, into a real incident. This skill finds those and seven other mistakes in apps that call LLMs before you ship, and tells you how to fix each one.
Your agent runs a scanner on your Python, JavaScript and TypeScript files (React, Vue and Svelte components included) that use OpenAI, Anthropic, Google GenAI, Mistral, Groq, Ollama, LangChain, LlamaIndex, the Vercel AI SDK or the OpenAI Agents SDK, and on your .env files. It skips comments and fixed strings, follows model output through variables, loops, .map() callbacks, React state and helper functions, and knows the usual fixes (query parameters, DOMPurify, an allow list of personas, needs_approval, a read-only database connection). It never runs your code or calls a model. It returns BLOCK, REVIEW, or NOT FLAGGED, with the file, line, reason, and a safer approach. Secrets are never printed.
What it checks
BLOCK
- L01: LLM API keys written in code (OpenAI, Anthropic, Groq, xAI, Hugging Face and other key formats, or literals passed to a client)
- L02: model output passed to eval, exec, subprocess, child_process or new Function
- L03: model output put into SQL text with f-strings, + or template literals
- L04: LLM keys sent to the browser (NEXT_PUBLIC_, VITE_ and EXPO_PUBLIC_ key variables, dangerouslyAllowBrowser)
REVIEW
- L11: request input in the system or developer prompt
- L12: model calls with no output token limit (Anthropic calls always have one and are not flagged)
- L13: full prompts, chat messages or model responses written to logs
- L14: model output rendered as raw HTML (dangerouslySetInnerHTML, innerHTML, v-html, unsafe_allow_html)
- L15: agent tools that run shell commands or code, or write files, without approval
- L16: model-written SQL run as it is, and text-to-SQL agents without a read-only connection
What's included
- The scanner (Python 3.9+, standard library only)
- A rules guide with the reason and safer approach for each rule
- A risky and a safe example project, and the real report the scanner produces on the risky one
Limits It reads each file on its own: model output passed to code in another file, prompt templates stored elsewhere, retrieved documents, provider-side settings and the sandbox your tools run in are not checked. NOT FLAGGED is not a guarantee.
Static check of source text, one file at a time. Data flow between files and runtime settings are not followed. NOT FLAGGED is not a guarantee.
How to use
- 1
Connect your agent to the Agensi MCP
Copy the prompt and paste it into Claude, ChatGPT, Cursor or any agent that supports MCP. It sets up the connection for you. Manual setup
- 2
Sign in to Agensi
Your agent shows a sign-in link. Sign in and click Allow.
- 3
Ask for what you need
Your agent loads this skill and uses it on your task. It always asks before any payment.
Reviews
No reviews yet
Be one of the first to try it. Every listed skill goes through the trust checks below.
Security scanned
Checked by our 8-point scan before listing
Fresh listing
Recently published to Agensi
Money-back guarantee
Doesn't do what its page promises? You get your money back.
Trust & safety
Security scan
Security scan passed 5 days ago
Permissions
Runs one local Python script in the terminal. It reads your project's files, writes nothing, and makes no network calls. The example project includes clearly fake, inert keys (not-a-real-... and AIzaNOTREAL...) for the scanner to find. It is never run.