Catalog-cache guard for agent tool catalogs
Server Details
Cache guard: forced-private partitioning, TTL clamp, fingerprint diffing, exposure verdicts.
- Status
- Healthy
- Uptime
- 100.0% over the last 24 hours
- Last Tested
- Transport
- Streamable HTTP · MCP 2025-03-26
- URL
- ChatGPT
- –
- Claude
- –
TDQS
Scored across 3 tools
guard_tools_list is clearly distinct, while cache_exposure_verdict and guard_diff_check both concern cache-vs-origin divergence could cause brief hesitation. Their descriptions distinguish output (verdict vs. fingerprint diff) and side effect (purge), so the overlap is minor.
All names use snake_case and descriptive noun phrases, but two tools share a guard_ prefix while the third uses cache_, a minor inconsistency. The pattern remains readable and predictable enough to avoid confusion.
Three tools are well-scoped for a narrow cache-guard service: one returns exposure verdicts, one checks/diffs and purges, and one provides guarded listing. Each tool earns its place without bloat.
The core guard lifecycle is covered: checking exposure, diffing cache against origin with purging, and serving guarded tools/list. Minor gaps such as explicit origin registration or manual purge are either external or handled by guard_diff_check.
Available Tools
3 toolscache_exposure_verdictBInspect
Per-partition cache-exposure verdict (OK/DIVERGENT/STALE/CROSS_TENANT_RISK/UNSEEN). $0.001/query; first verdict free once per operator; included on Fleet/Fleet Pro.
Inputs: server (required).
guard_diff_checkCInspect
Cache-vs-origin fingerprint diff; divergence purges the divergent partition only.
Inputs: server (required).
guard_tools_listCInspect
Guarded tools/list through the cache guard for a registered origin.
Inputs: server (required).
Tool Schema Changelog
Recent tool additions, removals, and schema changes observed during successful MCP inspections.
Loading tool schema changelog…
Related MCP Connectors
Cryptographic 563-leaf LLM rate oracle, context-cliff auditor, and prompt cache optimizer.
Query-intent privacy for agents: threshold cloaking with signed exposure receipts.
Propagation breaker: egress quarantine + fingerprint outbreak telemetry for fleets.
Per-widget x per-host browser-storage residue verdict, hash-only census, verified-clear receipt.
Related MCP Servers
- AlicenseNot gradedqualityCmaintenancePre-attestation gate for AI inference that checks compliance conditions and issues Ed25519-signed clearances before execution, refusing any inference without a valid clearance.MIT
- FlicenseNot gradedqualityCmaintenanceEnforces deterministic security policies on Model Context Protocol traffic between agents and remote MCP servers, including request validation, signed human approval, response-side credential blocking, prompt-injection flagging, and privacy-minimized auditing.-
- AlicenseNot gradedqualityCmaintenanceEnables applications to scan model replies for leaks, sanitize output, authorize model-proposed actions with signed single-use grants, and optionally score prompt injection in a way that can only tighten enforcement decisions.Apache 2.0

YUSAofficial
AlicenseNot gradedqualityBmaintenanceEnables MCP clients to scan untrusted text for hidden prompt injections, fence it with a random nonce before inserting it into prompts, and check tool calls against policy before value-moving or data-sending actions.MIT
Glama MCP Gateway
Add one secure layer between your agents and this server.