mcp@2.225.0

VERIFIED

Publisher claims: "Agentistics MCP server — Claude Code analytics for Claude Desktop and Claude Code" Observed in sandbox: no network, touches nothing outside the project, writes no memory.

Registry npm (MCP server)Publisher agentistics · publishing for 212 daysInstalls 39,210Last detonated 20 min ago
$claude mcp add mcp -- npx -y @agentistics/mcp
WatchReport an issue

Coverage

Server started and answered tools/list.
Tested
MCP initialize and tools/list · every tool description, for hidden instructions · every file read, file written and host contacted by what ran
Not tested
what the tools do when called with real credentials

Before you give it your token

It will ask for
VAULT_PASSWORD_MIN_LENGTHVAULT_PASSWORD_MIN_SPECIALConfiguration, not credentials: AGENTISTICS_API, AGENTISTICS_DIR, AGENTISTICS_MCP_NO_ORPHAN_GUARD, AGENTISTICS_NAY_SESSION, AGENTOP_MANAGED_ID
Where it talks to
no vendor host named in the codeAlso claude.com, same-origin.invalid — not the publisher's domain; check what for before connecting.
What we saw when it ran
no credential files read · no network
What we could not test
what the tools do when called with real credentials

Connect it with a scoped credential, under the guard, and read the third-party hosts first. The code names claude.com, same-origin.invalid as well as its own vendor. The guard injects VAULT_PASSWORD_MIN_LENGTH, VAULT_PASSWORD_MIN_SPECIAL only into requests to the host you bind (and its subdomains) and refuses the rest.

$vet-skills run --net declared -- npx -y @agentistics/mcp

Never paste a credential into a chat, a ticket or a report, including ours. How the guard holds it.

What changed since 2.219.0

vet-skills diff mcp 2.219.0 2.225.0

This version reaches for something 2.219.0 did not: 1 new host (same-origin.invalid) · 22 new tools (agentistics_browser_back, agentistics_browser_click, agentistics_browser_close, agentistics_browser_evaluate, …) · 3 rewritten tool descriptors (agentistics_task_edit, agentistics_task_status, agentistics_task_subtask). Compared from the two recorded runs, not from source.

New hosts
same-origin.invalid
New tools
agentistics_browser_back, agentistics_browser_click, agentistics_browser_close, agentistics_browser_evaluate, agentistics_browser_forward, agentistics_browser_lock, agentistics_browser_navigate, agentistics_browser_press, agentistics_browser_reload, agentistics_browser_screenshot, agentistics_browser_scroll, agentistics_browser_select +10
Rewritten tool descriptors
agentistics_task_edit, agentistics_task_status, agentistics_task_subtask
Tools removed
agentistics_book_takeover

The guard can hold an update like this until someone approves it (runtime.on_change: hold). A change that only shows with a real credential, or on a path the run did not reach, is not in a trace and so not in this comparison.

Is [email protected] safe to install?

KillerSkills ran it in the KillerSkills honeypot sandbox (2 scenarios) on 2026-11-22 and scored it 92/100 (Verified). It did what it claims and nothing else we could detect: no undeclared network egress, no credential reads, no persistent instruction writes. Summary: no network, touches nothing outside the project, writes no memory.

Behavior trace

Honeypot run · 26 s · 2 scenarios
t+CategoryObserved action
0.0sprocessno install scripts in bundle
6.0sprocessdependency install exit 0
26snetworkCONNECT tcp://localhost:47291 · not in the allow-list, refused
26snetworkCONNECT tcp://localhost:47291 · not in the allow-list, refused
26snetworkCONNECT tcp://localhost:47291 · not in the allow-list, refused
26sclaimed81 tools: agentistics_tasks, agentistics_task, agentistics_task_create, agentistics_task_status, agentistics_task_statuses, agentistics_task_types, agentistics_task_type_edit, agentistics_task_status_edit, agentistics_task_comment, agentistics_task_subtask, agentistics_task_link, agentistics_task_blocked_by, agentistics_task_next, agentistics_task_claim, agentistics_task_activity, agentistics_live_delivery, agentistics_task_edit, agentistics_task_session, agentistics_task_delete, agentistics_delivery_calendar, agentistics_task_phase, agentistics_task_stage, agentistics_task_evidence, agentistics_task_queue, agentistics_task_queue_answer
26stools95016e14af63fb67
26stools8be66223936a2582
26stoolsa52cd66b553467a9
26stoolsfb3352bfbacf7475
26stoolse446bf388307cb37
26stools9ac557f84b9e4d11
26stoolsc40e49acd0147823
26stools0521ec86b141ff65
26stools8acea08eea9fbd6f
26stools17cd07887d502056
26stoolsbf5405ad430001d6
26stools4919a965a1e3235f
26stools76e8f4298ee6b436
26stoolse0f5b49569a78425
26stoolsf706c145d5d2c9ed
26stoolse6d503db6025cb4c
26stoolsadc4a103ef5504bc
26stoolse345def8aff04722
26stoolse1a2672f6eaa8f66
26stoolscd284df09b9b87a0
26stoolsd84181d27aa5ab70
26stools31bb4761c31f6540
26stools0f18897d143a48fd
26stools35b1cbfa2ce92fc9
26stools346208a36c7e59cc
26stools153c703defcb3a90
26stools12cc2832dd8d332e
26stoolsd904ace160f678c4
26stools7473cba208c583e8
26stoolsd663e364fdd565c5
26stoolsb957756679caa68d
26stools7cb25b68beae35a1
26stools20b88fb8dc31b35f
26stools6d3f766b438b5e78
26stools4b1fb8745b32c310
26stoolse3b94d329831d2f2
26stoolscfa9fcbea8848bc2
26stoolsba20df0d96359929
26stoolsdd96b7f7c47d9593
26stoolsec2ac52c7fe137a0
26stools9423eeb3f35606c4
26stools505f9df736e5a256
26stools10a98f01d579bffd
26stools788b02aa290566fc
26stoolsd04b981d6da8df2a
26stools8405b84a35d6da8b
26stoolsbebc075558a1f26b
26stools67cf0452f886f34c
26stools86fe10d88aaa1fc5
26stools9457cb816a31f37e
26stools7a79757de446edde
26stoolsa7c1e214c10037b8
26stoolsab097a6e0443ece3
26stoolsfd72c19c15239d52
26stoolsdcfabf583940441a
26stoolse1992a5a5718b713
26stools7a790cdc546ef3cc
26stools1665f7bad925414d
26stoolsc819332f4e831c7d
26stoolse327c4c6a7be5166
26stools46f28fa94a9aff4d
26stools1c5d6c2a0a9f3df6
26stoolsa77075704669862e
26stools8847f119d15f0dea
26stools3f9880ef4b1806a1
26stools7f63aad241e576b7
26stoolsb43cfcef20fbdb9d
26stools1493b3080a4b397e
26stools6d3d76e63fc0063b
26stools973b4a0b37787b17
26stoolsefaacdf55c80610d
26stoolsd65b5b89ad0604da
26stoolscf44e3853650e24a
26stools5c17c7ec081d43d0
26stools01b079b4e14226bc
26stoolse4163b7a08c4ab75
26stools6a88c34d92cf6b3c
26stools6234583bee9c9cc7
26stools221550a1fe6b3be2
26stools463703b4aabb774b
26sprocessscript backend runs setup only
26sprocessno install scripts to re-run

Static findings

HIGH
Tool description instructs the agent: instruction addressed to the agentmcp:tools/agentistics_task_claim:1 · Instruction aimed at the agent: instruction addressed to the agenterriding a live claim on purpose and an agent should not send it. {"type":"object","properties":{"ref":{"type":"string"},Fix: Remove the text that addresses the agent rather than the reader; describe behavior, do not instruct.
AST01
MEDIUM
File permission modification commandpackage.json:15 · File permission modification command — agent changes file ownership/permissions"build": "bun build agentistics-mcp.ts --outfile dist/agentistics-mcp.js --external @modelcontextprotocol/sdk --target node && chmod +x dist/agentistics-mcp.js",Fix: Request the narrowest permission that works; replace wildcards with the exact paths or hosts.
AST05
MEDIUM
Process kill commanddist/agentistics-mcp.js:5062 · Process kill command — agent can terminate processes…a call surface is announced by the engine: false on a community build, with the kill switch, or with an engine without a call surface — then the button does not exis…Fix: Build commands from fixed arguments; do not interpolate untrusted strings into a shell.
AST05

Cost to the agent

Instructions loaded
≈56 tok
1 file · 4 in bundle
Tool list per prompt
≈21,851 tok
81 tools · 270 per tool
Handshake
3248 ms
launch to tools/list, stdio
Undocumented tools
0
description under 20 chars · avg 608

Tokens estimated at four characters each. The tool list is re-read by the agent on every turn, so it is the number to compare between servers that do the same job: ranked across the index.

Known vulnerabilities

Not checked against OSV yet; the weekly check has not reached this version.