AgentDesk MCP
@Rih0z
About AgentDesk MCP
Adversarial AI review API — independent AI reviews another AI's output. Stop LLMs from grading their own homework. Provides automated quality assurance for AI-generated code, content, and other outputs through independent review pipelines.
Config
Add this server to your MCP-compatible client using the configuration below.
{
"mcpServers": {
"agentdesk-mcp": {
"command": "npx",
"args": [
"-y",
"@ezark-publish/[email protected]"
]
}
}
}Tools
4Adversarial quality review of any AI-generated output. An independent reviewer assumes the author made mistakes and actively looks for problems. Returns structured verdict (PASS/FAIL/CONDITIONAL_PASS), score (0-100), categorized issues with severity, and evidence-based checklist. Works for any output type: code, content, summaries, translations, data extraction, etc.
Dual adversarial review: two independent reviewers assess the output from different angles, then a merge agent combines their findings. Stricter than single review — if either reviewer finds a critical issue, the merged verdict is FAIL. Use for high-stakes outputs where quality is critical.
List all available services on the AgentDesk marketplace. Returns service catalog with pricing, quality scores, and capabilities. Filter by category, minimum quality score, maximum price, or capability.
Execute a service on the AgentDesk marketplace. Requires an AgentDesk API key for authentication. Pass service-specific input parameters.
Overview
What is AgentDesk MCP?
AgentDesk MCP is an MCP tool that provides independent adversarial quality review for any AI-generated output. It integrates with Claude Code, Claude Desktop, and any MCP client, and uses your own Anthropic API key (BYOK) to perform structured, bias-resistant evaluations.
How to use AgentDesk MCP?
Install via npx agentdesk-mcp and add it to your MCP client, setting the ANTHROPIC_API_KEY environment variable. Two tools are provided: review_output for a single adversarial review, and review_dual for a dual‑reviewer pass with a merged verdict. Pass the AI output to review, optionally customizing criteria and review type.
Key features of AgentDesk MCP
- Adversarial prompting that assumes the author made mistakes
- Evidence-based checklist with automatic downgrade for missing evidence
- Anti‑gaming validation forces FAIL if >30% of items lack evidence
- Dual adversarial review with two independent reviewers
- One‑tool MCP setup, no SDK required
- BYOK (bring your own API key) — uses your own Anthropic API key
Use cases of AgentDesk MCP
- Code review: check for bugs, security issues, and performance problems
- Content review: verify accuracy, readability, SEO, and audience fit
- Factual verification: validate claims in AI-generated text
- Translation quality: assess accuracy and naturalness
- Data extraction: verify completeness and correctness of extracted data
FAQ from AgentDesk MCP
What makes adversarial review different from self‑review?
Self‑review has a systematic leniency bias because the same model shares blind spots that created errors. AgentDesk uses a separate reviewer invocation with adversarial prompting, which is fundamentally different and more rigorous.
How does dual review work?
Dual review (review_dual) runs two independent adversarial reviews from different angles, then a merge agent combines findings. If either reviewer finds a critical issue, the merged verdict is FAIL, and the lower score is taken. Issues are combined and deduplicated.
What are the limitations of AgentDesk MCP?
Like all LLM‑as‑judge systems, it can be vulnerable to prompt injection. The anti‑gaming validation layer mitigates superficial gaming, but determined adversarial inputs remain a challenge. Each review_output call makes 1 LLM API call; review_dual makes 3, so factor BYOK cost into your pipeline.
What are the system requirements?
You need an ANTHROPIC_API_KEY environment variable. The default reviewer model is Claude Sonnet 4‑6, but you can specify another model. No other SDK or dependencies are required.
How does AgentDesk MCP compare to other evaluation tools?
AgentDesk MCP offers one‑tool MCP setup, adversarial review, dual reviewer capability, and anti‑gaming validation — features not found in manual prompts, Braintrust, or DeepEval. It is MCP native and requires no SDK.
Frequently asked questions
What makes adversarial review different from self‑review?
Self‑review has a systematic leniency bias because the same model shares blind spots that created errors. AgentDesk uses a separate reviewer invocation with adversarial prompting, which is fundamentally different and more rigorous.
How does dual review work?
Dual review (`review_dual`) runs two independent adversarial reviews from different angles, then a merge agent combines findings. If either reviewer finds a critical issue, the merged verdict is FAIL, and the lower score is taken. Issues are combined and deduplicated.
What are the limitations of AgentDesk MCP?
Like all LLM‑as‑judge systems, it can be vulnerable to prompt injection. The anti‑gaming validation layer mitigates superficial gaming, but determined adversarial inputs remain a challenge. Each `review_output` call makes 1 LLM API call; `review_dual` makes 3, so factor BYOK cost into your pipeline.
What are the system requirements?
You need an `ANTHROPIC_API_KEY` environment variable. The default reviewer model is Claude Sonnet 4‑6, but you can specify another model. No other SDK or dependencies are required.
How does AgentDesk MCP compare to other evaluation tools?
AgentDesk MCP offers one‑tool MCP setup, adversarial review, dual reviewer capability, and anti‑gaming validation — features not found in manual prompts, Braintrust, or DeepEval. It is MCP native and requires no SDK.
Basic information
More Developer Tools MCP servers
SSH MCP Server
hypnosisSSH MCP server for AI agents: remote commands, file transfer, log search and server audits through OpenSSH.
Reelier
Maxime HouleAgents make claims. Reelier writes receipts — record an agent's tool-call workflow once, replay it deterministically at 0 tokens, and diff runs to catch drift.
Mobbin
MobbinMobbin MCP connects your AI agents to 600,000+ real product screens, so what they build starts with what already works.
Apiosk
obcraftAI-native payments for tools and APIs. Discover, pay for, execute, and publish monetized APIs directly from your agent, settled per call in USDC over x402, through the Model Context Protocol. payments · finance · x402 ·
Routebase
RoutebaseRoutebase keeps your API's design, docs, contract tests, mocks and monitoring in one living OpenAPI spec so they can't drift apart. Agents work on that same source of truth, under your team's permissions.
Comments