larksnap-fetch

No direct malware is evidenced in the provided fragment because it is primarily documentation, not the daemon/extension implementation. The major security concern is supply-chain risk from executing remotely fetched installer scripts directly (curl|bash and irm|iex) combined with the described creation of persistent services and the later execution of external download/processing binaries (yt-dlp/ffmpeg). The snippet’s claimed runtime protections (loopback binding, Origin/header checks, and HMAC anti-replay) cannot be validated without the underlying code, so this should be treated as elevated risk pending verification of install.sh/install.ps1 and daemon/bridge source.

SUSPICIOUS. The skill is broadly aligned with its stated purpose, and the documented data path is local rather than a hidden SaaS proxy. However, it depends on a third-party browser extension and bundled daemon that leverage logged-in browser sessions to read and edit Feishu content, which is a substantial trust boundary and medium-high security risk even without clear evidence of malicious intent.

This module does not show classic malware/backdoor behaviors (no eval/exec/persistence). However, it has high-impact security risks centered on untrusted Markdown image embedding: it can download arbitrary http/https URLs (broad SSRF-like reach depending on runtime networking controls) and can potentially read arbitrary local files due to missing baseDir containment checks for decoded src paths. Because it base64-embeds those bytes into the outgoing Markdown and then transmits the payload to a remote document service, a realistic abuse scenario is sensitive data exfiltration if an attacker can control mdFile content or image src values. Overall, treat this fragment as security-relevant and review/mitigate SSRF/LFI and data-handling paths.