Works with the AI Agents you already use
Claude Code
Cursor
Codex CLI
DeepSeek
Muse+22 moreClaude CodeCursorCodex CLIDeepSeekMuse
Dots
Manus
Grok
Grok Bot+18 agentsRequest Guard MCP
Turn observed website and API requests into explained risk assessments. Companion agent instructions for the free Request Guard MCP server: classify metadata, inspect signals, and keep enforcement under your control.
$6
Request Guard MCP
Example session with this skill installed
Use Request Guard MCP to assess this observed API request: POST /wp-login.php, user agent sqlmap, source IP 203.0.113.9 (a reserved test address). Explain the verdict and supporting signals. Do not send cookies, authorization headers, or a request body.
- Read your context and instructions
- Compiled the request guard mcp
Request assessment
Verified example: the released Windows Rust v1.0.0 binary was called over local stdio with only POST, /wp-login.php, self-declared user agent sqlmap, and documentation-only IP 203.0.113.9. No cookies, authorization headers or request body were supplied.
| Field | Result |
|---|---|
| Verdict | block |
| Confidence | medium |
| Threat category | scraping |
| Score | 1.0 |
| Signal | ua_scraper, value 1.0, weight 0.7 |
Calling explain with the original request metadata returned a recomputed explanation: the score was 100%, the verdict was block, and ua_scraper contributed 100%. It recommended considering a WAF/load-balancer block and reviewing matching blocklists. This explanation recomputes from supplied metadata; it is not a lookup of a stored decision.
Next step: review application logs and independent signals before changing policy. User-agent strings can be spoofed. This heuristic does not establish identity or block the request itself. Your application supplies observed requests and enforces its own decision. Core classify/explain worked without external Redis, PostgreSQL or MaxMind services in this verification.
This is a protocol-level example for the v1.1 skill update's target server, not certification of every agent or deployment. The original skill v1.0 targeted server v0.1.4.
Connects securely to your tools. The creator never sees your data.
About this skill
Understand why a request was flagged
Request Guard MCP gives your agent tools to classify observed inbound HTTP request metadata and explain the contributing signals. This companion skill guides the agent to choose the right tool, minimize the metadata it sends, and distinguish a heuristic verdict from proof of identity.
For example, assess POST /wp-login.php with user agent sqlmap, then explain its ua_scraper signal. Your website or API supplies the observed metadata and decides what action to enforce. The skill does not intercept traffic or block visitors.
What the $6 purchase includes
An agent-instruction SKILL.md ZIP, installed separately from the free GPL-3.0-or-later server. Purchasing the skill is optional; the companion instructions are also available in the server repository.
The original skill v1.0 targets experimental server v0.1.4 over authenticated HTTP POST. The submitted v1.1 update targets stable Rust server v1.0.0 and documents local stdio as well as authenticated HTTP POST. Check the approved package version before installing.
Choose your implementation
- Rust server, native releases and setup: stable v1.0.0, native downloads and npm installer.
- C#/.NET port: in development, with no official release yet. Consult its parity matrix; this skill does not certify .NET compatibility.
Redis, PostgreSQL and MaxMind data enable optional tools and may have operating costs. Core classification needs no external backend. The server does not provide SSE GET or OAuth discovery. It assesses inbound requests; it does not control an agent's outbound requests.
Find more skills like this on the Claude Skills Marketplace, the Codex Skills Marketplace and the Gemini Skills Marketplace, or browse Claude Skills for Security.
Instruction ZIP only; install the free server separately. Original skill v1.0 targets experimental v0.1.4 HTTP POST; the v1.1 update targets stable Rust v1.0.0 with stdio or authenticated HTTP POST. Check the approved version. No traffic interception, automatic blocking, outbound enforcement, SSE GET or OAuth discovery. Verdicts are heuristics, not identity proof or site authorization. Optional backends have separate setup and costs. The .NET port remains in development; full current-Rust parity and every client/platform are not certified.
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 buying a skill.
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 3 days ago
- Software
- Request Guard MCP
- Optional
- RedisPostgreSQLMaxMind data
Install the free server separately. Skill v1.1 targets Rust v1.0.0 via local stdio or HTTP POST; HTTP requires a unique bearer token. Legacy skill v1.0 targets v0.1.4 HTTP POST. Check the approved package version.
Permissions
The agent sends selected inbound request metadata to the buyer-configured Request Guard MCP server. Avoid sending request bodies, cookies, authorization headers, or personal data unless the buyer’s task requires it. The skill does not need access to the buyer’s filesystem.