dxh141130

Automate punching time in/out on WPS Time / NetTime (wpstime.com NetTime). Use for phrases like setup punchclock/configure punchclock/set up time clock, clock in/clock out, start break/end break, start lunch/end lunch, check status/status. Runs a Playwright flow, captures a screenshot, and replies with a brief confirmation.

What to consider before installing: - Platform and binaries: This skill requires a macOS host (it uses the 'security' Keychain CLI) and Node.js with the Playwright package available. Confirm the agent/gateway runs on macOS and that Node + Playwright are installed, or ask the author to provide an install spec (package.json or explicit instructions). - Avoid the chat-based setup: Do NOT use the 'chat wizard' option that asks you to type your password into chat. Use the interactive local setup (node ./scripts/setup.mjs) so credentials go directly into the macOS Keychain and are not recorded in chat/gateway logs. - Confirm dependencies and scope: Ask the publisher to declare required binaries (node, playwright, security) and an OS restriction (macOS). Request a package.json or install instructions so you can review and control what gets installed. - Verify runtime behavior: Inspect and/or run the scripts in a controlled environment first. The scripts take screenshots and capture page text (which may contain sensitive info). Ensure screenshots are handled safely when attached to chat channels. - Principle of least privilege: If you intend to allow autonomous invocation, ensure the agent's execution environment is restricted (so it cannot run arbitrary shell commands beyond what this skill needs) and monitor logs for unexpected behavior. If you cannot confirm the above, treat the skill as risky. At minimum, ask the author to (1) add an install spec and declared OS/binary requirements, (2) remove or disable the chat-based password collection, and (3) document what data is captured in screenshots/snippets and where attachments are posted.

Type: OpenClaw Skill Name: wps-punchclock Version: 1.0.1 The skill is classified as suspicious due to its direct use of `node:child_process.execFile` in `scripts/punchclock.mjs` and `scripts/setup.mjs` to execute the `security` command for macOS Keychain interaction. While this capability is plausibly needed for the stated purpose of securely managing credentials, direct shell execution is a high-risk behavior that prevents classification as benign. There is no clear evidence of intentional malicious behavior, data exfiltration beyond the stated purpose, or prompt injection attempts against the agent.