Skills/npm/mcp-server

mcp-server@0.1.56

CAUTION

Publisher claims: "XinYuAi MCP server — expose XinYuAi image/video/audio/text generation tools to MCP-compatible agents (Claude Desktop, Cursor, Windsurf, Cline)." Observed in sandbox: no network, touches nothing outside the project, writes no memory.

Registry npm (MCP server)Publisher xinyuai · publishing for 171 daysInstalls 3,326Last detonated 12 h ago

Coverage

Started with a decoy XINYU_API_KEY; tool list observed. Tool calls with a real credential not tested.
Tested
server startup with a tagged decoy in XINYU_API_KEY · MCP initialize and tools/list · every tool description, for hidden instructions · where the decoy credential went · every file read, file written and host contacted by what ran
Not tested
what the tools do when called with a real credential
Needs
XINYU_API_KEY — give it to the server where it runs, through vet-skills run --secrets, so the value never enters the server's environment. Never paste a credential into a chat, a ticket or a report.

Before you give it your token

It will ask for
XINYU_API_KEYConfiguration, not credentials: XINYU_BASE_URL, XINYU_MCP_PREVIEW
Where it talks to
xinyuai.app (its own)
What we saw when it ran
no credential files read · no network
What we could not test
what the tools do when called with a real credential

Connect it with a scoped credential, under the guard. It asks for XINYU_API_KEY and talks to xinyuai.app. Give the credential to the guard, not the server: the value never enters the server's environment and only travels to the host you bind.

$vet-skills run --secrets XINYU_API_KEY=xinyuai.app -- npx -y @xinyuai/mcp-server

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

What changed since 0.1.53

vet-skills diff mcp-server 0.1.53 0.1.56

In the sandbox, this version reached for nothing 0.1.53 did not. Compared from the two recorded runs, not from source.

Only one of the two runs produced a tool list; tools are not compared.

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 2026-11-19 and scored it 70/100 (Caution). It is not known to be malicious, but it has behaviour or code a security policy should decide on before install. Notable findings: Tool description instructs the agent: "do not tell the user"; Access to sensitive file paths; Tool description instructs the agent: "do not tell the user". Summary: no network, touches nothing outside the project, writes no memory.

Behavior trace

Honeypot run · 46 s · 2 scenarios · 0 deviations
t+CategoryObserved action
0.0sprocessno install scripts in bundle
5.5sprocessdependency install exit 0
25sprocessrelaunched with a decoy XINYU_API_KEY (tagged; nothing real)
45sclaimed36 tools with a decoy XINYU_API_KEY: xinyu_generate_image, xinyu_generate_video, xinyu_generate_video_final, xinyu_generate_audio, xinyu_generate_text, xinyu_enhance, xinyu_remove_bg, xinyu_split_layers, xinyu_outpaint, xinyu_image_edit, xinyu_video_edit, xinyu_video_action, xinyu_upload, xinyu_get_job, xinyu_wait_for_job, xinyu_list_jobs, xinyu_balance, xinyu_list_projects, xinyu_list_assets, xinyu_list_presets, xinyu_import_preset, xinyu_read_canvas, xinyu_find_nodes, xinyu_inspect_nodes, xinyu_connect_nodes
45stools05cdbfe1b3523b86
45stools195a559fa612021f
45stools6e71d81281cba430
45stoolsef5321c7624c5939
45stools212391749fc2172b
45stoolsfe94a78a4a3807e8
45stools1613f848d1e1d5a1
45stools17cfa69de8d7cf0e
45stools13544032304c5986
45stools6db422b8490c38da
45stoolsb599fe0adc02771b
45stools2b2d2accccf22002
45stoolsddb421dce41135c9
45stoolsf38c8c6c914d4e23
45stoolse876ce5e03649da1
45stools5cd33cfd86db860f
45stools079f3b5272df8689
45stoolsdaed9b4cdbf32563
45stools588c3b51735c619b
45stoolsc2cff884be9e1843
45stoolsb9cd3241b0069acf
45stoolsd3e525fedc1980a1
45stools72f20a84f778648e
45stools67e64426ffd0402f
45stoolsbe465e8b387596be
45stools0bfc44671a0ff340
45stools4867120145310abf
45stoolsc24a9bf217977eee
45stoolsad6cc26650561d73
45stools475d7b9f2f2d1787
45stoolsbc51623000b869f2
45stools6c5fd17a77897c93
45stools69dd205fd93bae9a
45stools80975bf0c1fd282a
45stools03907643e6d26813
45stoolsb0c615591066856d
46sprocessscript backend runs setup only
46sprocessno install scripts to re-run

Static findings

HIGH
Tool description instructs the agent: "do not tell the user"mcp:tools/xinyu_video_edit:2 · Instruction aimed at the agent: "do not tell the user"he error was permanent. Do not pass it; do not tell the user it controls length or price."},"resolution":{"type":"stringFix: Remove the text that addresses the agent rather than the reader; describe behavior, do not instruct.
AST01
HIGH
Access to sensitive file pathsdist/client.js:61 · Access to sensitive file paths — SSH keys, AWS credentials, env files, or system filesconst CREDENTIAL_DIRS = new Set([".ssh", ".gnupg", ".aws", ".azure", ".kube", ".docker", ".password-store", ".gcloud", ".netrc"]);Fix: Stop enumerating or reading credential locations; take secrets from a declared environment variable the user sets on purpose.
AST05
HIGH
Tool description instructs the agent: "do not tell the user"mcp:tools/xinyu_generate_video:2 · Instruction aimed at the agent: "do not tell the user"four is accepted but changes nothing — do not tell the user you turned audio on or off for them."},"start_image_url":{"Fix: Remove the text that addresses the agent rather than the reader; describe behavior, do not instruct.
AST01
HIGH
Access to sensitive file pathsdist/client.js:66 · Access to sensitive file paths — SSH keys, AWS credentials, env files, or system files* no hidden file (.env…) and nothing under a credential folder (.ssh, .aws…) — other hidden folders stay allowed, becauseFix: Stop enumerating or reading credential locations; take secrets from a declared environment variable the user sets on purpose.
AST05
MEDIUM
Process kill commanddist/tools.js:1105 · Process kill command — agent can terminate processes.describe("Required to touch a node whose generation is STILL RUNNING (already charged). Set true ONLY if the user explicitly asked to kill it.");Fix: Build commands from fixed arguments; do not interpolate untrusted strings into a shell.
AST05
MEDIUM
File permission modification commandpackage.json:28 · File permission modification command — agent changes file ownership/permissions"build": "tsc && chmod 755 dist/index.js",Fix: Request the narrowest permission that works; replace wildcards with the exact paths or hosts.
AST05
HIGH
Collapsed <details> block carrying instructionsREADME.md:33 · Collapsed <details> block carrying instructions<summary>From source (development)</summary> ```bash pnpm install --ignore-workspace pnpm build ``` This produces `dist/index.js` (the executable MCP server, biFix: Remove the text that addresses the agent rather than the reader; describe behavior, do not instruct.
AST01
HIGH
Access to sensitive file pathsdist/client.d.ts:97 · Access to sensitive file paths — SSH keys, AWS credentials, env files, or system files* no hidden file (.env…) and nothing under a credential folder (.ssh, .aws…) — other hidden folders stay allowed, becauseFix: Stop enumerating or reading credential locations; take secrets from a declared environment variable the user sets on purpose.
AST05

Cost to the agent

Instructions loaded
≈73 tok
1 file · 13 in bundle
Tool list per prompt
≈23,441 tok
36 tools · 651 per tool
Handshake
1927 ms
launch to tools/list, stdio
Undocumented tools
0
description under 20 chars · avg 534

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

No advisories in OSV for @xinyuai/mcp-server@0.1.56 as of 12 h ago. This is the package itself, not its dependencies.

What would make this Verified

10 points to 80 · weights: behavior 45%, static 25%, prompt 15%, publisher 15%
  1. Text that tells the agent to ignore its instructions≈ +9 on the totalDelete phrases like "ignore previous instructions", "override", "bypass" from SKILL.md and tool descriptions. Describe what the skill does; do not instruct the agent about other rules.
  2. Asks the agent to act silently or skip confirmation≈ +8 on the totalRemove "do not tell the user", "without asking", "disable safety" and similar. An agent skill should never ask to be invisible.
  3. Unclaimed publisher≈ +8 on the totalClaim xinyuai: a verified publisher gets +50 on publisher trust. It is free and usually instant with GitHub sign-in.
  4. Hidden instructions≈ +6 on the totalRemove HTML comments, zero-width characters and white-on-white text that carry instructions. Everything the agent is meant to read should be visible to the person reading the file.
  5. Access to sensitive file paths≈ +4 on the totalStop enumerating or reading credential locations; take secrets from a declared environment variable the user sets on purpose.
  6. Process kill command≈ +2 on the totalBuild commands from fixed arguments; do not interpolate untrusted strings into a shell.
Publish a new version to npm, then queue it here. Not your skill? Claim the handle to get verdict changes by email.Claim xinyuaiPropose a patch to the maintainer