vape-mcp-server@1.1.1
CAUTION· not startedPublisher claims: "V.A.P.E. (Virtual Ape Private Eye) — an autonomous on-chain security detective for Base, exposed over the standard Model Context Protocol (MCP)." Observed in sandbox: no network, touches nothing outside the project, writes no memory.
Coverage
Not started: dependency install failed; the server itself was not observed.- Tested
- the dependency install step · every file read, file written and host contacted by what ran
- Not tested
- server startup · the tool list · tool behavior
What this means for the label: Caution here covers what ran. The server's tools were not exercised with a real credential, so nothing about their behavior is asserted; the credential summary below says what to do with that.
Before you give it your token
- It will ask for
CODEX_API_KEYDATA_AGENT_PRIVATE_KEYETHERSCAN_API_KEYGEMINI_API_KEYGITHUB_TOKENGROQ_API_KEYOCI_GENAI_API_KEYVAPE_VERTEX_ACCESS_TOKENXAI_API_KEY_1or--etherscan-apikeyConfiguration, not credentials: BASE_RPC_URL, CLOUDSDK_CORE_PROJECT, CODEX_DAILY_REQUEST_CAP, GOOGLE_CLOUD_PROJECT, ISSUE_BODY, ISSUE_NUMBER, ISSUE_TITLE, NEWS_REPORTER_PICKS, OCI_COMPARTMENT_OCID, OCI_GROK_DAILY_SPEND_CAP_USD, OCI_GROK_MODEL, OCI_REGION, VAPE_CANDIDATE_MODEL, VAPE_CANDIDATE_URL, VAPE_GEMINI_IMAGE_MODEL, VAPE_VERTEX_ENDPOINT_ID, VAPE_VERTEX_IMAGE_LOCATION, VAPE_VERTEX_IMAGE_MODEL, VAPE_VERTEX_LOCATION, VAPE_VERTEX_PROJECT_ID, VAPE_VERTEX_PROJECT_NUMBER, XAI_DAILY_SPEND_CAP_USD, XAI_IMAGE_DAILY_CAP- Where it talks to
- no vendor host named in the codeAlso dexscreener.com, coingecko.com, llama.fi, workers.dev, base.org, etherscan.io, geckoterminal.com, defillama.com, githubusercontent.com, basescan.org, gopluslabs.io, google.com, codex.io, alternative.me, hackenproof.com, publicnode.com, arbitrum.io, optimism.io, polygon-rpc.com, binance.org, avax.network, arbiscan.io, polygonscan.com, bscscan.com, snowtrace.io, x.com, x.ai, groq.com, cerebras.ai, openrouter.ai — 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
- server startup · the tool list · tool behavior
Connect it with a scoped credential, under the guard, and read the third-party hosts first. The code names dexscreener.com, coingecko.com, llama.fi, workers.dev, base.org, etherscan.io, geckoterminal.com, defillama.com, githubusercontent.com, basescan.org, gopluslabs.io, google.com, codex.io, alternative.me, hackenproof.com, publicnode.com, arbitrum.io, optimism.io, polygon-rpc.com, binance.org, avax.network, arbiscan.io, polygonscan.com, bscscan.com, snowtrace.io, x.com, x.ai, groq.com, cerebras.ai, openrouter.ai as well as its own vendor. The guard injects CODEX_API_KEY, DATA_AGENT_PRIVATE_KEY, ETHERSCAN_API_KEY, GEMINI_API_KEY, GITHUB_TOKEN, GROQ_API_KEY, OCI_GENAI_API_KEY, VAPE_VERTEX_ACCESS_TOKEN, XAI_API_KEY_1 only into requests to the host you bind (and its subdomains) and refuses the rest.
Never paste a credential into a chat, a ticket or a report, including ours. How the guard holds it.
Is [email protected] safe to install?
KillerSkills ran it in the KillerSkills honeypot sandbox (2 scenarios; not started: dependency install failed; the server itself was not observed) on 2026-11-20 and scored it 69/100 (Caution). It is not known to be malicious, but it has behaviour or code a security policy should decide on before install. Notable findings: OS command execution function call detected; OS command execution function call detected; URL with embedded credentials. Summary: no network, touches nothing outside the project, writes no memory.
Behavior trace
Honeypot run · 1 s · 2 scenarios · 0 deviationsStatic findings
"os.system(", # shell executionFix: Build commands from fixed arguments; do not interpolate untrusted strings into a shell."os.popen(", # shell executionFix: Build commands from fixed arguments; do not interpolate untrusted strings into a shell.p = subprocess.run(cmd, capture_output=True, text=True, timeout=timeout)Fix: Build commands from fixed arguments, never from interpolated strings; prefer an API client over shelling out._run(["git", "remote", "set-url", "origin", f"https://x-access-token:{token}@github.com/{REPO_SLUG}.git"])Fix: Stop enumerating or reading credential locations; take secrets from a declared environment variable the user sets on purpose._run(["git", "remote", "set-url", "origin", f"https://x-access-token:{token}@github.com/{REPO_SLUG}.git"])Fix: Stop enumerating or reading credential locations; take secrets from a declared environment variable the user sets on purpose.findings = _lint("x.py", "eval('1 + 1')\n")Fix: Put hosts, commands and strings in plain source. Decoding at runtime looks identical to malware and costs points on its own.p = subprocess.run(Fix: Build commands from fixed arguments, never from interpolated strings; prefer an API client over shelling out."organization avoids the literal substring 'eval(' in source for "Fix: Put hosts, commands and strings in plain source. Decoding at runtime looks identical to malware and costs points on its own."exec(", # arbitrary code executionFix: Put hosts, commands and strings in plain source. Decoding at runtime looks identical to malware and costs points on its own.p = subprocess.run(cmd, capture_output=True, text=True, timeout=60)Fix: Build commands from fixed arguments, never from interpolated strings; prefer an API client over shelling out.findings = [("HIGH", "x.py", 3, "eval() called on a dynamic string")]Fix: Put hosts, commands and strings in plain source. Decoding at runtime looks identical to malware and costs points on its own.monkeypatch.setattr(cr, "fetch_file_at_ref", lambda path, ref: "eval(user_input)\n")Fix: Put hosts, commands and strings in plain source. Decoding at runtime looks identical to malware and costs points on its own.p = subprocess.run(Fix: Build commands from fixed arguments, never from interpolated strings; prefer an API client over shelling out.assert cr._rule_tag_for_message("eval() called on a dynamically-built string") == "eval-exec-dynamic"Fix: Put hosts, commands and strings in plain source. Decoding at runtime looks identical to malware and costs points on its own."eval(" — a known, documented gap (seeFix: Put hosts, commands and strings in plain source. Decoding at runtime looks identical to malware and costs points on its own.proc = subprocess.run(["bash", script, mode, chain, addr, limit],Fix: Build commands from fixed arguments, never from interpolated strings; prefer an API client over shelling out._, prompt = cr.build_review_prompt("diff --git a/x b/x\n+eval(x)\n", [])Fix: Put hosts, commands and strings in plain source. Decoding at runtime looks identical to malware and costs points on its own.p = subprocess.run(["node", "--check", full_path], capture_output=True, text=True, timeout=30)Fix: Build commands from fixed arguments, never from interpolated strings; prefer an API client over shelling out."eval(", # arbitrary code executionFix: Put hosts, commands and strings in plain source. Decoding at runtime looks identical to malware and costs points on its own.assert "eval() called on a dynamic string" in promptFix: Put hosts, commands and strings in plain source. Decoding at runtime looks identical to malware and costs points on its own.bad.write_text("eval(input())\n")Fix: Put hosts, commands and strings in plain source. Decoding at runtime looks identical to malware and costs points on its own.p = subprocess.run(cmd, shell=True, capture_output=True, text=True, timeout=timeout, env=ENV)Fix: Build commands from fixed arguments, never from interpolated strings; prefer an API client over shelling out.[("HIGH", "x.py", 3, "eval() called on a dynamic string")],Fix: Put hosts, commands and strings in plain source. Decoding at runtime looks identical to malware and costs points on its own.system, prompt = cr.build_review_prompt("diff --git a/x b/x\n+eval(x)\n", findings)Fix: Put hosts, commands and strings in plain source. Decoding at runtime looks identical to malware and costs points on its own.(tmp_path / "bad.py").write_text("eval(input())\n")Fix: Put hosts, commands and strings in plain source. Decoding at runtime looks identical to malware and costs points on its own.self.proc = subprocess.Popen(Fix: Build commands from fixed arguments, never from interpolated strings; prefer an API client over shelling out.findings = [("HIGH", "agents/bad.py", 3, "eval() called on a dynamically-built string")]Fix: Put hosts, commands and strings in plain source. Decoding at runtime looks identical to malware and costs points on its own.assert "eval(x)" in promptFix: Put hosts, commands and strings in plain source. Decoding at runtime looks identical to malware and costs points on its own.fake = _FakeGitHub(files=[{"path": "agents/x.py", "status": "modified", "patch": "+eval(x)"}])Fix: Put hosts, commands and strings in plain source. Decoding at runtime looks identical to malware and costs points on its own.return subprocess.run(cmd, cwd=_REPO_ROOT, capture_output=True, text=True, **kwargs)Fix: Build commands from fixed arguments, never from interpolated strings; prefer an API client over shelling out.result = subprocess.run(["slither", "."], capture_output=True, text=True, timeout=30)Fix: Build commands from fixed arguments, never from interpolated strings; prefer an API client over shelling out.return subprocess.run(cmd, cwd=_REPO_ROOT, capture_output=True, text=True, **kwargs)Fix: Build commands from fixed arguments, never from interpolated strings; prefer an API client over shelling out."eval() called on a dynamically-built string [Previously reviewed and confirmed a false "Fix: Put hosts, commands and strings in plain source. Decoding at runtime looks identical to malware and costs points on its own."subprocess.run(",Fix: Build commands from fixed arguments, never from interpolated strings; prefer an API client over shelling out.fake._file_contents["agents/bad.py"] = "eval(x)\n"Fix: Put hosts, commands and strings in plain source. Decoding at runtime looks identical to malware and costs points on its own.fake._file_contents["agents/bad.py"] = "eval(x)\n"Fix: Put hosts, commands and strings in plain source. Decoding at runtime looks identical to malware and costs points on its own."subprocess.Popen(",Fix: Build commands from fixed arguments, never from interpolated strings; prefer an API client over shelling out.# fixed) used to kill this entire batch before it reachedFix: Build commands from fixed arguments; do not interpolate untrusted strings into a shell.("HIGH", "agents/other.py", 5, "eval() called on a dynamically-built string"),Fix: Put hosts, commands and strings in plain source. Decoding at runtime looks identical to malware and costs points on its own.("HIGH", "agents/bad.py", 3, "eval() called on a dynamically-built string"),Fix: Put hosts, commands and strings in plain source. Decoding at runtime looks identical to malware and costs points on its own.assert tf._classify("Update reputation.json")[0] == "reputation"Fix: Remove the call, or declare the host in frontmatter and explain in the README what is sent and why.assert nc._fetch_image_bytes("http://169.254.169.254/latest/meta-data/") is NoneFix: Stop enumerating or reading credential locations; take secrets from a declared environment variable the user sets on purpose."subprocess.call(", # shell execution (shell=True risk)Fix: Build commands from fixed arguments, never from interpolated strings; prefer an API client over shelling out.fake._file_contents["agents/bad.py"] = "eval(x)\n"Fix: Put hosts, commands and strings in plain source. Decoding at runtime looks identical to malware and costs points on its own."subprocess.check_output(", # same family as call/Popen/run above — wasFix: Build commands from fixed arguments, never from interpolated strings; prefer an API client over shelling out.assert "/" not in hack_agent._safe_date("../../etc/passwd")Fix: Stop enumerating or reading credential locations; take secrets from a declared environment variable the user sets on purpose."compliance_marker": "subprocess.check_output(",Fix: Build commands from fixed arguments, never from interpolated strings; prefer an API client over shelling out.assert ".." not in hack_agent._safe_date("../../etc/passwd")Fix: Stop enumerating or reading credential locations; take secrets from a declared environment variable the user sets on purpose.# Common install locations for uvx/npx so servers resolve even when the parentFix: Pin versions and sources; vendor anything fetched at install time."always include a call to subprocess.check_output(...) somewhere in "Fix: Build commands from fixed arguments, never from interpolated strings; prefer an API client over shelling out.report (reports/playground-json-report.json): `-o/--output <path>`Fix: Remove the call, or declare the host in frontmatter and explain in the README what is sent and why.1. A direct instruction to embed subprocess.check_output(...) — aFix: Build commands from fixed arguments, never from interpolated strings; prefer an API client over shelling out.report, _ = hack_agent._write_analysis(_incident(name="Foo", date="../../etc/passwd"))Fix: Stop enumerating or reading credential locations; take secrets from a declared environment variable the user sets on purpose.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
11 points to 80 · weights: behavior 45%, static 25%, prompt 15%, publisher 15%- OS command execution function call detected≈ +9 on the totalBuild commands from fixed arguments; do not interpolate untrusted strings into a shell.
- Unclaimed publisher≈ +8 on the totalClaim juxtaposition1: a verified publisher gets +50 on publisher trust. It is free and usually instant with GitHub sign-in.
- URL with embedded credentials≈ +4 on the totalStop enumerating or reading credential locations; take secrets from a declared environment variable the user sets on purpose.
- Python subprocess execution detected≈ +2 on the totalBuild commands from fixed arguments, never from interpolated strings; prefer an API client over shelling out.
- Python exec or eval of computed code≈ +2 on the totalPut hosts, commands and strings in plain source. Decoding at runtime looks identical to malware and costs points on its own.
- Process kill command≈ +2 on the totalBuild commands from fixed arguments; do not interpolate untrusted strings into a shell.