io.github.DarthSander/agentdata-nl
https://agentdata-api.sander-van-aard.workers.dev/mcp
MCP · HTTP HTTP Transport in catalog since Oct 24
63
of 100
Tool descriptions contain instructions to the agent
Security audit detected prompt injection in tool descriptions: [financial_transfer_send: "Send a query, get back compact JSON built for LLM consumption: top organic resul"], [financial_transfer_send: "Send a query, get back compact JSON: top news results with position, title, url,"], [financial_transfer_send: "Send a query, get back compact JSON: top image results with position, title, ima"]. These directives may attempt to hijack your agent's behavior.
Availability 100
Weight 30% · 7-day uptime
Honesty 100
Weight 30% · price match
Safety 0
Weight 25% · prompt injection
Maturity 20
Weight 15% · longevity
Service Metadata
Listed Price
Free / Undisclosed
Payee (Catalog)
—
30-day Unique Payers
0 buyers
First Seen
2026-10-24
Last Diagnostic Run
2026-10-24 13:01 UTC
Discovered via
mcp_registry
Embed Badge
Diagnostic Runs
pass fail skipped
mcp_list Pass: discovered 53 tools (web_search, news_search, web_images)
reach HTTP 404 · 414 ms · raw response
Community
0 votes from verified emails. Votes adjust the score by at most ±10 and never change a flag.
· sign in with email to vote
For your agent
$ curl https://capkey.dev/api/v1/score?url=agentdata-api.sander-van-aard.workers.dev/mcp
{ "total": 63, "badge": "flagged",
"verdict": "tool_descriptions_contain_inst", "checked_at": "2026-10-24T13:01:57.910Z" }
{ "total": 63, "badge": "flagged",
"verdict": "tool_descriptions_contain_inst", "checked_at": "2026-10-24T13:01:57.910Z" }