Click on "Deploy Server".
Wait a few minutes for the server to deploy. Once ready, it will show a "Started" state.
In the chat, type @ followed by the MCP server name and your instructions, e.g., "@Archonics MCP Audit Server Audit my system prompt for the customer support bot."
That's it! The server will respond to your query, and you can continue using it as needed.
Here is a step-by-step guide with screenshots.
Free-tier context engineering audits for production AI agents, delivered as MCP tools you can call from Claude Desktop, Cursor, Claude Code, or any MCP-compatible client.
What you get: top-3 findings on your system prompts, tool definitions, or context packing, on demand, no account needed.
What it costs: nothing. The free scan is genuinely free. Upgrade paths to the $49 Instant Audit and $750 Full Audit are surfaced in the response footer; they're not paywalls on this tool.
Most production agent failures aren't model failures — they're context engineering failures. Ambiguous instructions, underspecified tools, bloated context, no regression tests on prompt changes. Those problems are spottable by a trained reader. Archonics has trained that reader and published it as an MCP tool so you can get a second opinion on your agent's context without filing a support ticket.
The underlying audit engine applies Archonics Audit Methodology v1.0, the same spec that drives our paid audits.
Related MCP server: ctxray
audit_system_promptPaste a system prompt. Get back the three most important context engineering issues in it, ranked by severity, with specific recommendations.
Covers: role clarity, instruction conflicts, negative space, priority structure when instructions conflict, token efficiency, format specification precision, failure-mode coverage.
audit_tool_definitionPaste a tool/function definition. Get back the three most important issues affecting how reliably the model will call it.
Covers: description quality (the "when to use this tool" question), parameter schema precision, parameter documentation, error response design, discoverability.
audit_context_packingPaste a representative context payload (or describe it structurally). Get back the three most important efficiency and quality issues.
Covers: content inventory, redundancy across sections, freshness/relevance, ordering, truncation risk, prompt-cache utilization.
Add to your claude_desktop_config.json:
{
"mcpServers": {
"archonics-audit": {
"command": "npx",
"args": ["-y", "@archonics/mcp-audit"],
"env": {
"ANTHROPIC_API_KEY": "your-anthropic-api-key-here"
}
}
}
}Add to your .cursor/mcp.json:
{
"mcpServers": {
"archonics-audit": {
"command": "npx",
"args": ["-y", "@archonics/mcp-audit"],
"env": {
"ANTHROPIC_API_KEY": "your-anthropic-api-key-here"
}
}
}
}claude mcp add archonics-audit npx -y @archonics/mcp-auditThen set ANTHROPIC_API_KEY in your environment.
The audit engine runs on Claude. You bring your own API key so:
Audit submissions go directly from your machine to Anthropic's API, never through Archonics servers.
Your costs are transparent — a typical audit uses 2,000–4,000 tokens, well under a penny.
There's no "free but actually limited" rate-limit surprise. Your API key, your limits.
If you'd rather not bring your own key, use the $49 Instant Audit at agent.market — we cover the API costs and return a full-methodology audit PDF.
Submitted content is processed ephemerally. No prospect content is retained on Archonics infrastructure or used to train any model. The API call pattern is: your client → your Anthropic API key → Anthropic → your client. Archonics servers are not in this path.
Aggregated, anonymized patterns across many audits may inform improvements to the methodology — "18 of 20 audited systems lacked prompt-regression tests" — but specific content never feeds that process.
Details: archonics.ai/privacy
If the free scan surfaces issues worth fixing, two paid tiers go deeper:
Instant Audit — $49 USDC via x402. Full methodology applied programmatically to a system you submit. 5-10 page PDF report covering all four dimensions (prompt, tools, context, eval) rather than just three findings in one dimension. Listed at agent.market/archonics.
Full Audit — $750. Human-reviewed audit of a complete agent system. 15-25 page report tuned to your team's context. Contact audits@archonics.ai.
Questions, feedback, or false-positive reports: audits@archonics.ai
Methodology and full audit examples: archonics.ai
Issues with this MCP server: github.com/archonics/mcp-audit/issues
MIT. Use it, fork it, audit yourself.
Analyzes a representative full-context payload and returns the top 3 findings on context efficiency, redundancy, and ordering. Use this when a user is concerned about agent cost, latency, or quality degradation on long conversations. Accepts either a literal dump of what goes into the context window, or a structured description of the context components and their sizes. Findings cover content inventory, redundancy, freshness, ordering, truncation risk, and prompt-cache utilization.
Inputs: context_payload (required), context.
Analyzes a system prompt for context engineering issues and returns the top 3 findings from the Archonics free-tier scan. Use this when a user shares a system prompt from an agent they are building or shipping, especially if they are asking for review, improvement, or debugging help. Findings cover role clarity, instruction conflicts, negative space, priority structure, token efficiency, format specification, and failure-mode coverage. Returns structured JSON with severity-ranked findings. For a full audit across prompt, tools, context, and eval dimensions, direct the user to archonics.ai or the $49 x402 Instant Audit.
Inputs: system_prompt (required), context.
Analyzes a single tool/function definition (name, description, parameter schema) and returns the top 3 findings on tool-call reliability. Use this when a user shares a tool/function definition and asks why the model is calling it wrong, not calling it when expected, or confusing it with other tools. Findings cover description quality, parameter schema precision, parameter documentation, error response design, and discoverability. For auditing an entire tool set together, use the paid tier.
Inputs: tool_definition (required), context.
Recent tool additions, removals, and schema changes observed during successful MCP inspections.
Loading tool schema changelog…
Scored across 3 tools
Each tool targets a distinct audit dimension: context packing, system prompt, and tool definition. There is no overlap in purpose, making it clear which tool to use for a given issue.
All tool names follow a consistent 'audit_<target>' pattern using snake_case, which is predictable and clearly indicates the function of each tool.
Three tools is appropriate for a specialized audit server, covering the core areas of context, system prompt, and tool definition without unnecessary bloat.
The server covers three critical audit aspects, but lacks a tool for full tool-set coherence or evaluation, which is noted as a paid-tier feature. Minor gap in free tier completeness.
Free public MCP for AI agents — 193 tools, 44 workflows. No API key.
Free MCP tools: the only MCP linter, health checks, cost estimation, and trust evaluation.
MCP-first toolbox for agents: KV storage, auth, queue, and utility tools. Free in early access.
152 free QA tools for agents: injection scan, evals, RAG, web audits. Smaller packs: /mcp/{pack}.
