mitracore — trust layer for agent work
One call to verify agent output, price its resource use and enforce spending mandates
- Pricing
- Freemium
- Platforms
- API, CLI, self-hosted
- Runs
- Remotely
- Version
- 0.2.0
- Listing updated
- 3 days ago
About mitracore — trust layer for agent work
mitracore — trust layer for agent work is a freemium MCP server for security. It runs as a hosted remote server, so there is nothing to install locally.
Mitracore bundles three checks that an agent workflow usually needs separately. Verification inspects an artifact in four layers: schema conformance, emptiness such as uniform lists, stub values and truncation, generation provenance, and authority, meaning whether the action was allowed. It returns an Ed25519-signed receipt that can be verified offline.
The quote tool prices what a call consumed, from tokens and CPU seconds to bytes out, storage writes and external calls, against a published rate card with the breakdown attached. A mandates feature covers limits, allowlists and a kill switch. A companion guard proxy sits in front of a paying agent and decides before money moves, applying a daily budget, per-call cap and payee and network allowlists, and recording each allow or deny in a signed, chained journal. Its decisions are made locally.
Pricing is posted on the site: 25 free calls a day, then $0.02 a call in USDC on Base over x402, or $19 for 30 days without a daily cap. The code is MIT licensed, with English and Hindi interfaces.
Install
Run in a terminal
claude mcp add --transport http mitracore https://rta.lat/mitracore/mcpAdd to Settings → Connectors → Add custom connector
https://rta.lat/mitracore/mcpAdd to ~/.cursor/mcp.json
{
"mcpServers": {
"mitracore": {
"url": "https://rta.lat/mitracore/mcp"
}
}
}Add to .vscode/mcp.json
{
"servers": {
"mitracore": {
"type": "http",
"url": "https://rta.lat/mitracore/mcp"
}
}
}Add to ~/.codeium/windsurf/mcp_config.json
{
"mcpServers": {
"mitracore": {
"serverUrl": "https://rta.lat/mitracore/mcp"
}
}
}Add to your client's MCP settings (most use this shape)
{
"mcpServers": {
"mitracore": {
"type": "http",
"url": "https://rta.lat/mitracore/mcp"
}
}
}Tools this server provides (5)
Read directly from the server.
- verify_work
- Verify an AI agent artifact in four layers: schema, emptiness, provenance and mandate authority. Returns an attest verdict.
- check_skill
- Privately vet a SKILL.md: structure, triggers, leaked secrets, prompt-injection markers and broken links. Returns a verdict, a score and a sha256 ready to be signed.
- quote_cost
- Compute the price from measured usage — tokens, CPU seconds, bytes, storage writes and external calls — instead of a flat price table.
- trust_report
- One verdict for an AI agent artifact: work verification (attest) + price from measured usage (agentpay) + mandate availability (drift).
- suite_status
- Live status of the Suite layers: attest, agentpay, drift, skilllint.
What it does
- Four-layer verification
- Checks schema, emptiness, provenance and authority of an artifact.
- Signed receipts
- Ed25519 receipts that can be verified offline.
- Cost quoting
- Prices tokens, CPU, bytes out, storage writes and external calls from a rate card.
- Spending guard
- Local proxy enforces budgets, caps, allowlists and a kill switch.
- Chained journal
- Every allow or deny decision is signed and chained.
- Skill linting
- Includes a tool for checking an agent skill.
Good for
- Check an agent's output is not empty or truncated before using it
- Put a hard daily budget in front of an agent that pays per call
- Get an itemised cost for a single agent call
- Keep tamper-evident records of agent spending decisions
Pricing
Freemium, from $19 a month
25 free calls per day; $0.02 per call in USDC, or $19 for 30 days
Prices change. Check rta.lat before you buy.
mitracore — trust layer for agent work FAQ
What does it cost?
The site lists 25 free calls a day, $0.02 a call over x402, or $19 for 30 days with no daily cap.
Do I need an account?
No. Payment is a signed USDC request from your own wallet, with no account or API key.
Can receipts be checked without trusting the service?
Yes. The README says receipts are Ed25519-signed and verifiable offline.
What licence is it under?
MIT.
Is mitracore — trust layer for agent work free?
mitracore — trust layer for agent work has a free plan, with paid plans for more usage or features. Paid plans start at $19 a month.
How do I add mitracore — trust layer for agent work to Claude Code?
Run this command in your terminal: claude mcp add --transport http mitracore https://rta.lat/mitracore/mcp
What are the best alternatives to mitracore — trust layer for agent work?
The closest alternatives to mitracore — trust layer for agent work are REA, Codebase Memory, GitHub, Serena MCP: the IDE for your agent and XERJ.
Alternatives to mitracore — trust layer for agent work

REA
MCP serverCLI and MCP server that lets coding agents reverse engineer apps and binaries
Runs locally- Open source
- 464K GitHub stars
- Developer tools

Codebase Memory
MCP serverLocal MCP server that builds a code knowledge graph for AI coding agents
Runs locally- Open source
- 48.7K GitHub stars
- Knowledge and memory

GitHub
MCP serverOfficial MCP server connecting AI assistants to GitHub repos, issues and workflows
Not checked yet- Remote
- Open source
- 34.5K GitHub stars
- Developer tools

Serena MCP: the IDE for your agent
MCP serverMCP toolkit giving coding agents symbol-level code retrieval and editing
Runs locally- Open source
- 31.5K GitHub stars
- Developer tools

XERJ
MCP serverLocal search engine for AI agents with cited answers over your code and docs
Runs locally- Open source
- 19.2K GitHub stars
- Knowledge and memory

Ouroboros
MCP serverSpec-first layer that pins requirements and verifies results of AI coding agent work
Runs locally- Open source
- 6.4K GitHub stars
- Developer tools