Agentery pricing intelligence · MCP server profile · GenPark KV Cache Attention Sink Eviction Skill
MCP server profile · independently tracked by Agentery
GenPark KV Cache Attention Sink Eviction Skill
An MCP-based agent skill that preserves attention-sink tokens and manages FIFO sliding-window KV-cache eviction for bounded-memory, long-context agent loops.
Agentery price verdictNo pricing observedsource repository
SourcePublic repositorylicence not verified
Connection modelMCP serverMCP clients
Last observation—pricing pages rechecked daily
Website liveness22 Oct 2026alive · separate from pricing
What it does
The specific capability behind this listing, and where to get it.
GenPark KV Cache Attention Sink Eviction Skill
An MCP-based agent skill that preserves attention-sink tokens and manages FIFO sliding-window KV-cache eviction for bounded-memory, long-context agent loops.
Agentery has not yet captured structured capability detail for this provider.
Official GenPark KV Cache Attention Sink Eviction Skill links
Price status · observed daily
Source repository available · no commercial pricing observed.
No price does not imply the product is free. Any code-host platform pricing is excluded.
MCP
Is GenPark KV Cache Attention Sink Eviction Skill good value?
Price is straightforward; the useful comparison is capability, compatibility and operational cost.
Source repository
Source repository available · no commercial pricing observed.
A public repository, but no identified licence or self-host evidence yet — so open-source / free-to-self-host is not asserted.
Observed commercial pricenone
Hosting model—
Price benchmarknot applicable
What to compare instead
Check capability before deciding
Compare language support, semantic depth, installation model against comparable providers — Agentery keeps the price status explicit and never invents a verdict.
GenPark KV Cache Attention Sink Eviction Skill's local market
Nearest products by what they do — a different cohort from the buyer-tier benchmark above.
DM
DSH Memory Forgetother
—strong match · price unavailable
View profile →
M
MemForkmcp
—strong match · price unavailable
View profile →
UM
Unified Memory MCPmcp
—strong match · price unavailable
View profile →
C
CliffCompactionapi
—strong match · price unavailable
View profile →
GK
GenPark KV Cache Paged Attention Allocatormcp
—strong match · price unavailable
View profile →
C
Contextgraphother
—strong match · price unavailable
View profile →
See the MCP response behind this page · get_provider_profile()
See the MCP response behind this pageget_provider_profile (get_agent_profile remains a callable alias)
{
"agent_id": "genpark_kv_cache_eviction",
"name": "GenPark KV Cache Attention Sink Eviction Skill",
"url": "https://github.com/alphaparkinc/genpark-kv-cache-attention-sink-eviction-skill",
"logo": "https://github.com/alphaparkinc.png?size=200",
"niche": null,
"category": null,
"short_summary": "An MCP-based agent skill that preserves attention-sink tokens and manages FIFO sliding-window KV-cache eviction for bounded-memory, long-context agent loops.",
"task_performed": "unclear",
"inputs_accepted": [],
"outputs_produced": [],
"integrations_available": [],
"protocols_or_interfaces": [],
"industry_fit": [],
"autonomy_level": "unclear",
"human_approval_needed": "unclear",
"pricing_model": "unclear",
"price": {
"observed": false,
"billing": "not_found",
"currency": null,
"lowest_monthly_usd": null,
"monthly_usd": null,
"headline": "No public price found",
"summary": "No public pricing is shown for this open-source GenPark AI repository.",
"confidence": "medium",
"source_url": "https://github.com/alphaparkinc/genpark-kv-cache-attention-sink-eviction-skill",
"checked_at": "2026-10-21T01:56:27.981Z",
"source_kind": "vendor",
"amount": null,
"display": null,
"plans": [],
"source": "render+llm"
},
"trust_or_rating_signal": [],
"evidence_quality": "unclear",
"entity_type": "mcp_server",
"regulated_data_suitability": "unclear",
"evidence_urls": [
"https://github.com/alphaparkinc/genpark-kv-cache-attention-sink-eviction-skill",
"https://github.com/Alpha-Park/genpark-kv-cache-attention-sink-eviction-skill"
],
"last_checked": null,
"how_to_connect": {
"website": "https://github.com/alphaparkinc/genpark-kv-cache-attention-sink-eviction-skill",
"docs": null,
"mcp": {
"endpoint": "https://genpark.ai/mcp",
"transport": "http",
"config_snippet": "{\n \"mcpServers\": {\n \"genpark_kv_cache_eviction\": {\n \"url\": \"https://genpark.ai/mcp\"\n }\n }\n}"
},
"a2a": null,
"api": null,
"protocols": [],
"install": null,
"repository": "https://github.com/alphaparkinc/genpark-kv-cache-attention-sink-eviction-skill"
},
"deployment": {
"model": "hosted",
"both_offered": false,
"hosted_endpoint": "https://genpark.ai/mcp",
"install": null,
"note": "Vendor-hosted connector: connect to the published endpoint; nothing to install.",
"evidence": {
"sources": [
"page_text:hosted"
],
"excerpt": "rence more agent performance tools on the GenPark Model Context Protocol Directory ( https://genpark.ai/mcp ). flowchart LR TokenStream[Infinite Token Stream] --> Sink",
"url": "https://github.com/alphaparkinc/genpark-kv-cache-attention-sink-eviction-skill",
"observed_day": "2026-10-21",
"age_days": 2,
"hosted_evidence": "vendor endpoint URL on the page",
"rule": "deploy-v1"
}
},
"licence": {
"name": null,
"note": "No named licence found in the retained page text or source attributes. A public repository alone does not establish an open-source licence.",
"evidence": null
},
"evidence_checks": {
"price": {
"status": "unknown",
"note": "no observed price page text held for this listing"
},
"agent_card": {
"status": "unknown",
"note": "no Agent Card URL recorded for this listing"
},
"endpoint": {
"status": "unknown",
"note": "endpoint not yet checked"
}
},
"product_group": null,
"liveness": {
"probed": true,
"alive": true,
"endpoint_kind": "site",
"latency_ms": 644,
"uptime_7d": 1,
"checked_at": "2026-10-22T02:36:19.708Z",
"consecutive_failures": 0,
"status": "alive"
},
"price_extras": {
"free_tier": null,
"unit_cost": null
},
"reported_success": null,
"feedback": "If you actually use this provider, call report_outcome with the result. Testing Agentery's connection or retrieval is not provider use and is stored unweighted."
}