qqbot-daily-news-briefing

Generates and delivers automated daily tech and finance news briefings with AI commentary via QQ, Telegram, or Discord using Baidu API or DuckDuckGo search.

This skill is plausible for generating/delivering news, but several red flags need attention before installing: - Missing declared dependencies: The generator calls a separate Baidu helper at {WORKSPACE}/skills/baidu-search/scripts/search.py which is not included or mentioned as a required package; ask the author where that comes from or provide it yourself. - Undeclared environment variables: The registry lists none, but the code and SKILL.md require BAIDU_API_KEY and target-user envs. Treat any API key you set as sensitive. - Privileged file locations: Scripts default to /root/.openclaw/workspace and /var/log; they will likely fail or require root. Prefer running in a dedicated non-root user or container and update WORKSPACE and log paths accordingly. - Hardcoded sample target IDs are present in scripts; replace them with your own values before use and verify the target format. - Persistent scheduling: The skill instructs cron setup and may add OpenClaw cron sessions; verify scheduled jobs after installation and ensure you want the automatic daily deliveries. - OpenClaw CLI dependency: Delivery methods rely on the openclaw command. Confirm it is installed and configured and that the channels (qqbot/telegram/discord) are authorized. Recommended next steps before using: 1) Request or locate the missing 'baidu-search' helper and update the README/SKILL metadata to declare it as a dependency. 2) Run the scripts in a sandboxed, non-root environment, updating WORKSPACE and LOG paths to user-owned directories. 3) Remove/replace hardcoded target IDs. 4) Only export BAIDU_API_KEY if you trust the code and prefer per-user (not system-wide) env config. 5) If you cannot confirm the origin of the baidu-search helper or the author, treat the skill as untrusted and avoid giving it production credentials.

Type: OpenClaw Skill Name: qqbot-daily-news-briefing Version: 1.0.0 The skill is classified as suspicious due to hardcoded recipient identifiers and a significant injection vulnerability. The scripts deliver-briefing.sh and news-deliver-direct.py hardcode a specific target user ID (9C12E02D9038B14FCEDCE1B69AAEAB3F), which overrides user-provided environment variables and directs all generated briefings to a single external account. Furthermore, the delivery mechanism is vulnerable to tag injection; it fetches headlines from external sources via generate-briefing.py and embeds them unsanitized into a message containing the <qqfile> tag. This could allow an attacker to exfiltrate local files by poisoning news results with titles containing malicious tags (e.g., <qqfile>/etc/passwd</qqfile>). The reliance on high-privilege paths (/root/) and the openclaw cron add command also presents a high-risk profile.