barkpush
PassAudited by ClawScan on May 1, 2026.
Overview
This appears to be a legitimate Bark notification skill, but it handles push device keys, sends message content to the configured Bark API, and stores local push history.
Before installing, make sure the Bark endpoint is one you trust, protect config and history files, remove sample credentials, and manually review commands that push to all users or use critical/call/update/delete options.
Findings (4)
Artifact-based informational review of SKILL.md, metadata, install specs, static scan signals, and capability signals. ClawScan does not execute the skill or run runtime probes.
A mistaken or overbroad command could send disruptive notifications to many devices or alter prior pushed messages.
The skill can send to all configured users, issue critical notifications, and update or delete prior pushes. These are purpose-aligned and documented, but they are actions a user should intentionally approve.
支持全员推送：`--user all` ... `--level critical` ... `--delete abc123 --user alice`
Use explicit recipients where possible, review all-user and critical/call commands before sending, and keep update/delete operations tied to the intended push ID.
Anyone with the configured device keys could send notifications to those Bark devices.
Bark device keys are used as delegated push credentials and are included in outbound push requests.
single_payload["device_key"] = device_key ... resp = self._api.push_json(single_payload)
Store config files and environment variables securely, use only trusted Bark endpoints, and rotate device keys if they are exposed.
Local history may reveal private notification contents and push credentials to anyone who can read the state files.
The history schema shows that local history records may store recipient device keys and push message bodies.
"device_keys": { "description": "接收推送的设备密钥列表" } ... "body": { "description": "推送消息的正文内容" }Protect the .bark-push directory, keep history limits modest, and clear history if it contains sensitive messages or keys.
Users have less provenance information for deciding whether to trust the bundled code.
The artifacts include runnable Python files and CLI documentation, but the source provenance and installation mechanism are not fully specified.
Source: unknown ... No install spec — this is an instruction-only skill ... 15 code file(s)
Review the included code and prefer installing from a known, trusted repository or package source if available.