pushary@1.9.219
CAUTION· not startedPublisher claims: "Human-in-the-loop (HITL) for AI agents: ask a person for approvals or answers by push notification. One command connects Claude Code, Codex, Cursor, Gemini CLI and more." 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
PUSHARY_API_KEYor--no-ask-password/--with-tokenConfiguration, not credentials: ANSWER_PATH_SUFFIX, APPROVALS_PATH, BUILD_TAG, CLAUDE_CONFIG_DIR, CLIENT_REPORT_PATH, CODEX_EVENT_KEY, CODEX_HOME, DEVICE_KEY_PATH, DISPLAY, EXECUTOR_NUMBER, GROK_SESSION_ID, HERMES_COMMAND_ALLOWLIST_KEY, HERMES_GATE_TOOLS_KEY, HERMES_HOME, IDENTITY_PATH, JENKINS_HOME, MACHINES_PATH, OPENCODE_CONFIG_DIR, PATHEXT, ProgramFiles, ProgramW6432, PUSHARY_ADMISSION_FILE, PUSHARY_AGENT_NAME, PUSHARY_API_URL, PUSHARY_APP_REGISTER_TIMEOUT_MS, PUSHARY_AUTO_DECISION_BEACON, PUSHARY_AUTO_NOTIFY_SESSION_END, PUSHARY_BASE_URL, PUSHARY_BELL_QUIET, PUSHARY_BELL_UPGRADE_AT, PUSHARY_CLAUDE_BIN, PUSHARY_COMMAND_DIR, PUSHARY_CONFIG_FILE, PUSHARY_GATE_TIMEOUT_MS, PUSHARY_MCP_PATH_PREFIX, PUSHARY_MODEL_PRICING, PUSHARY_NO_BROWSER, PUSHARY_RECEIPTS, PUSHARY_SPAWN_ID- Where it talks to
- pushary.com (its own)Also nodejs.org — 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 nodejs.org as well as its own vendor. The guard injects PUSHARY_API_KEY only into requests to pushary.com (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.
What changed since 1.9.150
vet-skills diff pushary 1.9.150 1.9.219In the sandbox, this version reached for nothing 1.9.150 did not. Compared from the two recorded runs, not from source.
- No longer contacted
- apple.com, claude.ai, discord.gg, microsoft.com
1.9.150 was not run in the sandbox (install failed); the comparison covers what static analysis recorded only.
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; not started: dependency install failed; the server itself was not observed) on 2026-11-20 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: Auto-update or dynamic fetch configuration; Auto-update or dynamic fetch configuration; Auto-update or dynamic fetch configuration. Summary: no network, touches nothing outside the project, writes no memory.
Behavior trace
Honeypot run · 2 s · 2 scenarios · 0 deviationsStatic findings
const offReason = autoUpdateOffReason(input);Fix: Pin versions and sources; vendor anything fetched at install time."\`npx pushary@latest setup\` with the key Hermes uses, or Session history turned on in the Pushary Mac app")Fix: Pin versions and sources; vendor anything fetched at install time.if (mode?.kill || mode?.mode === "terminal_only") return;Fix: Build commands from fixed arguments; do not interpolate untrusted strings into a shell.…ray(value)) return false; const snapshot = value; return typeof snapshot.hadMcpServers === "boolean" && typeof snapshot.mcpServersWasRecord === "boolean" && (!snapshot.hadMcpServers || snapshot.mcpServersWasRecord || Object.hasOwn(snapshot…Fix: Namespace tool names and keep tool descriptions literal; no instructions to the agent inside them.var autoUpdateState = () => join7(pusharyDir(), "auto-update.json");Fix: Pin versions and sources; vendor anything fetched at install time.console.log(` ${dim("A browser on this computer is subscribed, but no phone is connected yet.")}`);Fix: Remove the call, or declare the host in frontmatter and explain in the README what is sent and why.result = subprocess.run([*helper, "authorize-export"],Fix: Build commands from fixed arguments, never from interpolated strings; prefer an API client over shelling out.var autoUpdateOffReason = (input) => {Fix: Pin versions and sources; vendor anything fetched at install time.// `npx ... doctor` reports npx for that invocation and claude_plugin asFix: Pin versions and sources; vendor anything fetched at install time.result = subprocess.run(command, input=json.dumps(payload), text=True, env=_registration_environment(key),Fix: Build commands from fixed arguments, never from interpolated strings; prefer an API client over shelling out.…ugins/pushary` and selects Pushary as the approval transport. On any computer, `npx pushary@latest setup --agents hermes` does the same. The plugin can also come from the Hermes …Fix: Pin versions and sources; vendor anything fetched at install time.…n**Example - dangerous command approval:**\n\n```json\n{\n "question": "Allow: rm -rf /tmp/build-artifacts/*",\n "type": "confirm",\n "context": "Cleaning up 2.3GB o…Fix: Build commands from fixed arguments; do not interpolate untrusted strings into a shell.var ENVIRONMENT_FILE_NAME = ".env";Fix: Stop enumerating or reading credential locations; take secrets from a declared environment variable the user sets on purpose.console.log(` ${dim('This computer keeps its own key. Revoke the extra "Paired agent (mobile)" key at')} ${cyan("pushary.com/dashboard/agent/settings")}${dim(".")}`);Fix: Remove the call, or declare the host in frontmatter and explain in the README what is sent and why.if (checkedWithinInterval(input.state.lastCheckAt, input.now, autoUpdateCheckInterval(input.state))) return { kind: "skip", reason: "checked_recently" };Fix: Pin versions and sources; vendor anything fetched at install time.…ools (grep, list_dir, glob, read_file, todo_write, search_tool) only honour the kill switch and a deny rule, as Claude Code's reads do. Grok fails open if it kills t…Fix: Build commands from fixed arguments; do not interpolate untrusted strings into a shell.return { ...unknownMode(), kill: last.state.kill === true };Fix: Build commands from fixed arguments; do not interpolate untrusted strings into a shell.return `no: it ${STOPS_AT_LOGOUT}; run \`pushary daemon install\`, or \`sudo loginctl enable-linger $USER\``;Fix: Build commands from fixed arguments; do not interpolate untrusted strings into a shell.text: `Phone-start ${STOPS_AT_LOGOUT}. To keep it running: \`sudo loginctl enable-linger ${outcome.user}\``Fix: Build commands from fixed arguments; do not interpolate untrusted strings into a shell.var autoUpdateDecision = (input) => {Fix: Pin versions and sources; vendor anything fetched at install time.if (mode?.kill || mode?.mode === "terminal_only") return;Fix: Build commands from fixed arguments; do not interpolate untrusted strings into a shell.var autoUpdateReportState = (decision) => {Fix: Pin versions and sources; vendor anything fetched at install time.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
No advisories in OSV for pushary@1.9.219 as of 40 min 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%- Unclaimed publisher≈ +8 on the totalClaim github actions: a verified publisher gets +50 on publisher trust. It is free and usually instant with GitHub sign-in.
- Auto-update or dynamic fetch configuration≈ +4 on the totalPin versions and sources; vendor anything fetched at install time.
- Destructive file system command detected≈ +4 on the totalBuild commands from fixed arguments; do not interpolate untrusted strings into a shell.
- 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.
- Privileged sudo command≈ +4 on the totalBuild commands from fixed arguments; do not interpolate untrusted strings into a shell.
- npx invocation of non-well-known package≈ +2 on the totalPin versions and sources; vendor anything fetched at install time.