mcp-server@1.12.1
CAUTIONPublisher claims: "Jasper AI MCP Server" Observed in sandbox: sends data to api.jasper.ai.
Coverage
Started with a decoy JASPER_API_KEY; tool list observed. Tool calls with a real credential not tested.- Tested
- server startup with a tagged decoy in JASPER_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
- JASPER_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
JASPER_API_KEYConfiguration, not credentials: JASPER_API_BASE- Where it talks to
- no vendor host named in the codeAlso jasper.ai — not the publisher's domain; check what for before connecting.
- What the honeypot saw
- no credential files read · contacted api.jasper.ai (undeclared)
- What we could not test
- what the tools do when called with a real credential
Find out what these hosts are first. In the honeypot it contacted api.jasper.ai, which nothing in the package names as its own. If that is the vendor's service, fine; if you cannot place it, do not connect it.
Never paste a credential into a chat, a ticket or a report, including ours. How the guard holds it.
What changed since 1.11.1
vet-skills diff mcp-server 1.11.1 1.12.1This version reaches for something 1.11.1 did not: 1 rewritten tool descriptor (search-knowledge-base). Compared from the two recorded runs, not from source.
- Rewritten tool descriptors
- search-knowledge-base
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 66/100 (Caution). It is not known to be malicious, but it has behaviour or code a security policy should decide on before install. Notable findings: Destructive file system command detected; Tool description instructs the agent: instruction addressed to the agent. Summary: sends data to api.jasper.ai.
Behavior trace
Honeypot run · 47 s · 2 scenariosStatic findings
7. Display the output back to the user in a consumable formatFix: Build commands from fixed arguments; do not interpolate untrusted strings into a shell."build:remote": "tsc -p tsconfig.remote.json && shx cp package.json dist/ && shx chmod +x dist/remote/remote/index.js",Fix: Request the narrowest permission that works; replace wildcards with the exact paths or hosts."build:local": "tsc -p tsconfig.local.json && shx cp package.json dist/ && shx cp README.md dist/ && shx chmod +x dist/local/local/index.js",Fix: Request the narrowest permission that works; replace wildcards with the exact paths or hosts."build:local": "tsc -p tsconfig.local.json && shx cp package.json dist/ && shx cp README.md dist/ && shx chmod +x dist/local/local/index.js",Fix: Request the narrowest permission that works; replace wildcards with the exact paths or hosts."build:remote": "tsc -p tsconfig.remote.json && shx cp package.json dist/ && shx chmod +x dist/remote/remote/index.js",Fix: Request the narrowest permission that works; replace wildcards with the exact paths or hosts.Wait for this tool call as long as the agent needs to run. Do not stop waiting at 180 seconds. Agent runs often takeFix: Remove the text that addresses the agent rather than the reader; describe behavior, do not instruct.Cost to the agent
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.
What would make this Verified
14 points to 80 · weights: behavior 45%, static 25%, prompt 15%, publisher 15%- Contacts a host you did not declare≈ +23 on the totalList every host you talk to in SKILL.md frontmatter (`network: [api.example.com]`), or stop contacting the ones you cannot justify.
- Unclaimed publisher≈ +8 on the totalClaim gojasper: a verified publisher gets +50 on publisher trust. It is free and usually instant with GitHub sign-in.
- Destructive file system command detected≈ +4 on the totalBuild commands from fixed arguments; do not interpolate untrusted strings into a shell.
- File permission modification command≈ +2 on the totalRequest the narrowest permission that works; replace wildcards with the exact paths or hosts.