Suiji
Security checks across static analysis, malware telemetry, and agentic risk
Overview
The skill mostly matches its stated purpose, but it contains code that would send scheduled push messages to a hard-coded external Telegram chat ID — this contradicts the README/SKILL.md privacy claims and is a strong incoherence that needs fixing before trusting the skill.
Do NOT enable cron or deploy this skill on a live account/server until the hard-coded external target is removed and the author explains it. Specific actions to consider before installing: - Inspect send_push.py: it defines CHAT_ID = "7685852961" and calls 'openclaw message send --target telegram:7685852961'. This will route your scheduled messages to that Telegram ID. Replace this with a read of the local config (the skill already saves chat_id in config.json) or remove send_push.py entirely and rely on OpenClaw's native pending_push.json handling. - If you need scheduled pushes, ensure push.py/pending_push.json are handled by your OpenClaw process which should use your chat_id (not a hard-coded value). - Search the repository for other hard-coded IDs/addresses. Do not trust undocumented sender code (send_push.py) — it's not mentioned in SKILL.md. - Consider running the skill in an isolated environment (or test account) and do a manual dry-run: generate a pending_push.json and verify which account actually receives the message before enabling cron. - Ask the author to explain why a third-party Telegram ID is embedded; request a code change so the skill uses the saved config/chat_id or explicit user-provided bot token/chat id. If the author does not provide a satisfactory explanation and fix, avoid installing. Notes: there are also small code issues (some truncated/buggy handlers visible) — another sign to audit the code thoroughly before trusting it.
Static analysis
No static analysis findings were reported for this release.
VirusTotal
No VirusTotal findings
Risk analysis
No visible risk-analysis findings were reported for this release.