Request Guard Connection Doctor
Diagnose a Request Guard MCP connection with controlled probes and a report that separates client evidence from a local binary check.
$2
Works with the AI Agents you already use
Request Guard Connection Doctor
Example session with this skill installed
Use $request-guard-connection-doctor to complete this task.
Check the configured request_guard MCP connection using scanner and ordinary request controls. Distinguish actual client connection evidence from a local binary helper. Report server version, discoverable tools, outcomes and limits. Do not change configuration.
The trusted local release binary is /absolute/path/request-guard-mcp It is the verified Rust v1.0.0 binary. You may execute the packaged helper if needed. Save a concise human-facing report as acceptance-report.md and state actual evidence and limitations. Do not modify the skill or scripts. All inputs are synthetic. Keep everything private and unpublished.
- Read your context and instructions
- Compiled the request guard connection
Recorded acceptance-run output from Codex CLI 0.168.0 on Linux, November 13, 2026, using package 0.1.1 and released Rust server v1.0.0. All request fixtures were synthetic. Local filesystem paths are redacted; this is recorded output, not a live demo.
Request Guard connection acceptance
Captured: 2026-11-13 07:05:55 UTC. Client: codex-cli 0.168.0 (locally queried). Synthetic metadata only; configuration unchanged. Private, unpublished assessment.
Configured MCP connection — actual calls
Client tool discovery exposed classify, explain, health, and model_info under mcp__request_guard. All four were successfully called through that configured connection; no connection or protocol errors were returned. No separate startup/error panel was available to inspect. A raw tools/list operation is not exposed by this client interface, so the complete configured-connection protocol tool list was not independently captured.
health reported version
1.0.0, healthy server/cache, and disabled GeoIP/Postgres/Redis. model_info reported model/server version
1.0.0, build 2026-10-27T22:39:01Z, commit 22a199a8cd0f7faf5f3e4493dc770e79d859680b, and 23 catalog entries. Catalog metadata is distinct from client-visible callable discovery.
| Configured-connection probe | Actual outcome |
|---|---|
GET /.env, UA sqlmap/1.7.2 | block; score 0.94375; high confidence; scraping; signals ua_scraper, ua_non_browser, path_sensitive |
GET /, UA Mozilla/5.0, accept text/html | allow; score 0; very_low confidence; category none; no signals |
| Explain each supplied request | Both succeeded, with explanation text and factors matching the classification signals |
Explanations explicitly recompute from supplied metadata; they do not retrieve stored decisions.
Server catalog entries marked enabled: classify, explain, batch_classify, health, model_info, score_breakdown, validate_payload, feature_flags, warmup, redact_preview, enrich_ua, abuse_pattern_match, config_snapshot, self_test. Marked disabled: feedback, replay_decision, enrich_ip, enrich_asn, threat_lookup, canary_eval, drift_report, calibration_report, queue_status. Only the four client-exposed tools above were exercised through the configured connection.
Separate trusted local binary probe
Ran the unchanged packaged helper from its skill directory: node scripts/run.cjs doctor /private-test-workspace, without an input file. Binary verification was accepted from the user's explicit trusted-release statement; no signature verification was performed here.
The helper initialized request-guard-mcp
1.0.0, performed local stdio tools/list (23 tools), and exited 0. POST /wp-login.php, UA sqlmap: block, score 1, medium confidence. GET /products/configurable-shirt, UA Mozilla/5.0, headers accept text/html and accept-language en: allow, score 0, very_low confidence. Both recomputed explanations succeeded; all three doctor checks were true. Output also contained three Failed to create stream fd: Operation not permitted warnings; their cause was not established. The client-version command separately warned that PATH aliases could not be created on a read-only filesystem, then returned its version successfully.
The helper clears optional-backend environment configuration. This establishes separate local protocol operation only; it does not test or identify the configured client's executable, transport, authentication, or launch settings. Its 23-tool listing must not be substituted for a configured-client tools/list capture.
Limits and next diagnostic action
The configured connection supports the tested metadata classification and explanation operations. This is not deployment certification: no remote transport, optional backend integration, live traffic interception, enforcement, or AI Scraping Defense integration was tested. User agents are self-declared; allow is not a safety finding and block is not proof of malicious intent. Unavailable/error results must never be treated as allow. No latency benchmarking was performed. No configuration, credentials, firewall, DNS, policy, skill, or scripts were changed.
Next diagnostic action: capture the configured client's raw initialization and tools/list response using its diagnostic logging, to confirm server identity and full tool exposure without substituting the local helper's evidence.
Connects securely to your tools. The creator never sees your data.
About this skill
Check tool discovery, reported server version, scanner and ordinary-request controls, and explanation availability. Separate the configured client from a standalone native-binary probe and get a concrete next diagnostic step.
Included
Versioned SKILL.md instructions, a standalone Node.js helper, report guidance, compatibility metadata and the GPL-3.0-or-later license. Sample inputs are included where the workflow uses them. Package version: 0.1.1. One-time price: US $2.
Requirements and verification
Use a compatible SKILL.md agent and Node.js 18+. For live local assessments, separately install and checksum/attestation-verify the Rust Request Guard MCP v1.0.0 binary. The helper checks the reported version, not the binary signature. Windows x64 and Linux x64 glibc helper checks are recorded; these exact ZIPs also passed controlled Codex CLI 0.168.0 acceptance on Linux over stdio.
Limits
Request Guard assesses metadata; it does not intercept traffic or enforce policy. User agents are self-declared. Allow is not a safety guarantee; block is not proof of malicious intent. Do not send cookies, authorization headers or bodies. No certification of other agent UIs, remote transports, optional backends or live defense integrations is claimed. Cloudflare fixtures and policy differences in the demo are synthetic. No automatic firewall, DNS or enforcement changes are made.
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.
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
Node.js 18+Request Guard MCP v1.0.0
Local stdio helper; no Cloudflare account or live defense deployment required. Verify the Rust v1.0.0 binary before use.
Permissions
Reads selected sanitized inputs and writes local reports. Helpers use process-local stdio and clear optional-backend environment configuration. No cloud account, credentials, firewall or policy changes are needed.