agent-host-mcp@0.2.2
VERIFIED· install onlyPublisher claims: "Connect Codex sessions, remote execution, and project knowledge to MCP clients." Observed in sandbox: no network, touches nothing outside the project, writes no memory.
Coverage
Bundled scripts ran with no arguments; no agent task scenario for this bundle.- Tested
- 8 of 58 bundled scripts run with no arguments (the rest were not run: per-bundle cap), with decoy credentials in the environment · every file read, file written and host contacted by what ran
- Not tested
- use of the skill by an agent
What this means for the label: Verified 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
MCP_TOKENConfiguration, not credentials: AGENT_HOST_MCP_HOME, AGENT_MCP_CODEX_BIN, AGENT_MCP_DB, AGENT_MCP_PROJECT_DIR, AGENT_MCP_TEST_INLINE, CODEX_HOME- Where it talks to
- no vendor host named in the code
- What we saw when it ran
- no credential files read · no network
- What we could not test
- use of the skill by an agent
Connect it with a scoped credential, under the guard. It asks for MCP_TOKEN and talks to its vendor. 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. What the tools do with a real credential was not observed.
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; bundled scripts ran with no arguments; no agent task scenario for this bundle) on 2026-11-22 and scored it 84/100 (Verified). In what ran, nothing we could detect went wrong: no undeclared network egress, no credential reads, no persistent instruction writes. What it does with a real credential was not observed; the static findings and the publisher carry the rest of the score. Summary: no network, touches nothing outside the project, writes no memory.
Behavior trace
Honeypot run · 1 s · 2 scenarios · 0 deviationsStatic findings
subprocess.run(['npm', 'ci', '--omit=dev', '--ignore-scripts', '--no-audit', '--no-fund'],Fix: Build commands from fixed arguments, never from interpolated strings; prefer an API client over shelling out.process = subprocess.run(source.lstrip()[1:], shell=True, executable='/bin/sh', stdin=subprocess.DEVNULL)Fix: Build commands from fixed arguments, never from interpolated strings; prefer an API client over shelling out.result = subprocess.run(['node', str(Path(__file__).with_name('openwiki_ingest.mjs')),Fix: Build commands from fixed arguments, never from interpolated strings; prefer an API client over shelling out.raise ValueError("Set a valid --api_key, config api_key, or MCP_TOKEN in .env/the environment before starting agent-host-mcp.")Fix: Stop enumerating or reading credential locations; take secrets from a declared environment variable the user sets on purpose.subprocess.run(["node", str(Path(__file__).with_name("openwiki_export.mjs")),Fix: Build commands from fixed arguments, never from interpolated strings; prefer an API client over shelling out.subprocess.run([self.executable(), "-S", screen_id, "-X", "quit"], env=self.environment(),Fix: Build commands from fixed arguments, never from interpolated strings; prefer an API client over shelling out.value = eval(compile(ast.Expression(expression.value), '<execute>', 'eval'), namespace)Fix: Put hosts, commands and strings in plain source. Decoding at runtime looks identical to malware and costs points on its own.os.execvpe(sys.argv[2], sys.argv[2:], os.environ)Fix: Build commands from fixed arguments, never from interpolated strings; prefer an API client over shelling out.return subprocess.run(['git', '-c', 'core.hooksPath=/dev/null', *arguments], cwd=workspace,Fix: Build commands from fixed arguments, never from interpolated strings; prefer an API client over shelling out.process = subprocess.run([binary, "-C", directory, action], stdin=subprocess.DEVNULL,Fix: Build commands from fixed arguments, never from interpolated strings; prefer an API client over shelling out.result = subprocess.run(Fix: Build commands from fixed arguments, never from interpolated strings; prefer an API client over shelling out.result = subprocess.run([self.executable(), "-S", screen_id, "-Q", "select", "."],Fix: Build commands from fixed arguments, never from interpolated strings; prefer an API client over shelling out.result = subprocess.run(command, env=self.environment(), stdin=subprocess.DEVNULL,Fix: Build commands from fixed arguments, never from interpolated strings; prefer an API client over shelling out.exec(compile(tree, '<execute>', 'exec'), namespace)Fix: Put hosts, commands and strings in plain source. Decoding at runtime looks identical to malware and costs points on its own.subprocess.run(command, input="Authorization: Bearer " + os.environ["MCP_TOKEN"] + "\n",Fix: Build commands from fixed arguments, never from interpolated strings; prefer an API client over shelling out."""Explicit API key wins, then exported MCP_TOKEN, then working-directory .env."""Fix: Stop enumerating or reading credential locations; take secrets from a declared environment variable the user sets on purpose.token = dotenv_values(env_file or Path.cwd() / ".env", interpolate=False).get("MCP_TOKEN")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. Skills load their instruction text when triggered; the figure is what that costs.
Known vulnerabilities
Not checked against OSV yet; the weekly check has not reached this version.