twitter-automation

CRITICAL E005: Suspicious download URL detected in skill instructions.

• Suspicious download URL detected (high risk: 0.70). These URLs point to a single small-domain distribution (inference.sh, cli.inference.sh, dist.inference.sh, cloud.inference.sh) that offers a direct installer (curl | sh) and downloadable binaries — a common, higher-risk pattern because executables served from a non‑widely‑established domain can deliver malware even if checksums are published (they are useful but not sufficient without detached signatures or trusted provenance); the image and docs links themselves are low-risk, but the CLI/binary distribution is potentially unsafe unless you can verify the publisher and signatures out-of-band.

MEDIUM W011: Third-party content exposure detected (indirect prompt injection risk).

• Third-party content exposure detected (high risk: 0.90). The SKILL.md explicitly includes apps like x/post-get and x/user-get and examples that fetch/read Twitter/X user-generated tweets and profiles via the X API, and those untrusted third-party contents can be read and used to drive actions (likes, retweets, follows, DMs), enabling indirect instruction injection.

MEDIUM W012: Unverifiable external dependency detected (runtime URL that controls agent).

• Potentially malicious external URL detected (high risk: 0.90). The skill's Quick Start directs running "curl -fsSL https://cli.inference.sh | sh", which fetches and executes a remote install script (from cli.inference.sh / dist.inference.sh) as a required dependency for the infsh CLI, so remote code is executed at setup/runtime.