Daily Morning Meeting

每日早会自动执行技能。多Agent抓取全网热点，整理成早会简报发送给CEO。

This skill appears to do what it claims (fetch news, format a briefing, and send it), but it has sloppy/incoherent elements you should resolve before installing on production systems. Specifically: - Confirm provenance: the Source/Homepage are unknown; prefer skills from a trusted author or with a public repo. - Inspect/modify save_path: both scripts hard-code C:\Users\Admin\.qclaw\workspace\ — change this to a correct, intended path for your environment to avoid accidental writes or collisions (and update the SKILL.md example for non-Windows hosts). - Verify the 'message' tool and recipient: ensure the message tool is authorized to send on behalf of the agent and that 'CEO' maps to the intended account; otherwise the skill could send data unexpectedly. - Review dependencies: SKILL.md lists qclaw-cron-skill and agent-orchestrator, but the included code does not invoke an orchestrator. If you plan to use cron/agent orchestration, prefer wiring scheduling/orchestration externally rather than trusting undocumented internal behavior. - Shell invocation hygiene: the Python script builds commands and calls subprocess.run(shell=True). If you adapt the skill to accept user-provided inputs (queries/recipients), ensure inputs are sanitized or avoid shell=True to prevent command injection. - Test in a sandbox: run the skill in an isolated environment, verify it only calls the expected openclaw tools, check what the message tool transmits (media file contents), and confirm no unexpected endpoints are contacted. If you can't verify the author's identity or the behavior of the dependent tools (web_search, message, agent-orchestrator), treat this as untrusted and avoid granting it access to real CEO communications or sensitive environments.

Type: OpenClaw Skill Name: daily-morning-meeting Version: 1.0.1 The skill contains a shell injection vulnerability in main.py due to the use of subprocess.run(shell=True) to execute tool calls with string-formatted parameters. Additionally, both main.py and run.js use hardcoded absolute Windows file paths (C:\Users\Admin\...), which is a poor security practice and limits portability. While the logic appears aligned with the stated purpose of news aggregation, the use of unsafe execution methods and hardcoded environment assumptions warrants a suspicious classification.