
Agent Infra Mcp
- Updated March 25, 2026
- lulzasaur9192/agent-infra-mcp
io.github.lulzasaur9192/agent-infra-mcp is an MCP server that exposes 18 tools for agent memory, API drift detection, and HMAC-signed trust receipts.
About
io.github.lulzasaur9192/agent-infra-mcp is a Model Context Protocol server that wires eighteen tools into Claude Code, Cursor, and similar agents for operational agent infrastructure. developers running multi-step automations often lose context between runs, miss silent API breaking changes, and have no portable audit trail when a tool call mattered for compliance or debugging. This server connects to three backed services—agent memory, drift detection, and trust receipts—each authenticated with its own environment variable. After registration, your agent can persist decisions, compare live API behavior against expectations, and attach HMAC-signed receipts to sensitive actions. It fits developers shipping agent-first products or internal bots who need production-grade hooks without building custom observability from scratch. Complexity is intermediate because you must provision three API keys and understand when to record memory versus emit a receipt. It is an MCP integration layer, not a replacement for your application’s primary database or full SIEM stack.
- 18 MCP tools across agent memory, API drift detection, and HMAC-signed trust receipts
- Stdio npm package @lulzasaur9192/agent-infra-mcp (server version 1.0.2)
- Separate API keys for Agent Memory, Agent Drift Detection, and Trust Receipt services
- Designed for long-running agent stacks that must prove what changed and what was trusted
- GitHub source at lulzasaur9192/agent-infra-mcp
Agent Infra Mcp by the numbers
- Data as of Jul 7, 2026 (Skillselion catalog sync)
claude mcp add --env AGENT_MEMORY_API_KEY=YOUR_AGENT_MEMORY_API_KEY --env AGENT_DRIFT_API_KEY=YOUR_AGENT_DRIFT_API_KEY --env TRUST_RECEIPT_API_KEY=YOUR_TRUST_RECEIPT_API_KEY agent-infra-mcp -- npx -y @lulzasaur9192/agent-infra-mcpAdd your badge
Show developers this MCP server is listed on Skillselion. Paste this into your README.
| Package | @lulzasaur9192/agent-infra-mcp |
|---|---|
| Transport | STDIO |
| Auth | None |
| Last updated | March 25, 2026 |
| Repository | lulzasaur9192/agent-infra-mcp ↗ |
What it does
Give your coding agent persistent memory, API drift alerts, and cryptographically signed trust receipts while you ship and run agent workflows.
Who is it for?
Best when you're operating Claude Code or Cursor agents that call external APIs repeatedly and need memory plus drift and receipt tooling in one MCP bundle.
Skip if: Skip if you only need a simple notes file or a full enterprise security operations center without third-party agent-infra API accounts.
What you get
After install and API keys, your agent can persist memory, flag drift, and produce HMAC-signed trust receipts on demand inside the same chat workflow.
- Agent-accessible memory persistence across tool calls
- Drift detection signals when external APIs diverge from expected behavior
- HMAC-signed trust receipts for selected agent actions
By the numbers
- 18 MCP tools documented in the server description
- 3 environment variables: AGENT_MEMORY_API_KEY, AGENT_DRIFT_API_KEY, TRUST_RECEIPT_API_KEY
- npm package @lulzasaur9192/agent-infra-mcp at server schema version 1.0.2
README.md
@lulzasaur9192/agent-infra-mcp
MCP server with 18 tools for AI agent infrastructure — persistent memory, API drift detection, and HMAC-signed trust receipts.
Installation
npx @lulzasaur9192/agent-infra-mcp
Claude Desktop / Cursor
{
"mcpServers": {
"agent-infra": {
"command": "npx",
"args": ["-y", "@lulzasaur9192/agent-infra-mcp"],
"env": {
"AGENT_MEMORY_API_KEY": "your-key",
"AGENT_DRIFT_API_KEY": "your-key",
"TRUST_RECEIPT_API_KEY": "your-key"
}
}
}
}
Tools (18)
Memory (8 tools)
Persistent key-value store with namespaces, metadata, TTL, and FTS5 full-text search.
memory_store— Store a memory (namespace + key + value)memory_store_batch— Batch store up to 100 memoriesmemory_list— List memories in a namespacememory_get— Get a specific memorymemory_update— Update value, metadata, or TTLmemory_delete— Delete a memory or entire namespacememory_search— Full-text search with metadata filteringmemory_stats— Usage statistics
Drift Detection (7 tools)
Monitor API endpoints for unexpected changes with golden test cases and cron scheduling.
drift_create_test— Register a golden test (exact/subset/contains/status match)drift_list_tests— List all testsdrift_run_test— Execute a test manuallydrift_test_history— View run history and diffsdrift_update_test— Update test configurationdrift_delete_test— Remove a testdrift_dashboard— Overview with drift rate and response times
Trust Receipts (3 tools)
HMAC-SHA256 signed, tamper-proof records of agent actions for audit trails.
receipt_issue— Issue a signed receipt for an agent actionreceipt_verify— Verify receipt integrity (online or offline)receipt_list— List receipts with filters
API Keys
Each service uses its own API key:
| Service | Env Var | API |
|---|---|---|
| Memory | AGENT_MEMORY_API_KEY |
agent-memory-api |
| Drift | AGENT_DRIFT_API_KEY |
agent-drift-api |
| Receipts | TRUST_RECEIPT_API_KEY |
trust-receipt-api |
License
MIT
Recommended MCP Servers
How it compares
MCP-backed agent operations toolkit, not a single Claude skill or generic logging CLI.
FAQ
Who is io.github.lulzasaur9192/agent-infra-mcp for?
It is for and developers who run coding agents in production-like loops and want memory, drift checks, and signed trust receipts without building custom infra.
When should I use io.github.lulzasaur9192/agent-infra-mcp?
Use it during Operate and late Ship when agents call live APIs, you care about contract drift, and you need auditable receipts after automated changes.
How do I add io.github.lulzasaur9192/agent-infra-mcp to my agent?
Install the npm stdio package @lulzasaur9192/agent-infra-mcp, set AGENT_MEMORY_API_KEY, AGENT_DRIFT_API_KEY, and TRUST_RECEIPT_API_KEY, then register the server in your Claude Code or Cursor MCP config.