autocli-web-scraping

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

• Third-party content exposure detected (high risk: 1.00). The skill's SKILL.md explicitly includes browser-based scraping commands for public social sites (e.g., autocli twitter, reddit, youtube), a declarative adapter pipeline with a "fetch" step that pulls arbitrary URLs, and integration guidance for AI agents to run those commands (including social-media actions), so the agent will ingest untrusted, user-generated web content that can influence follow-up actions.

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

• Potentially malicious external URL detected (high risk: 1.00). The one-line installer uses curl -fsSL https://raw.githubusercontent.com/nashsu/autocli/main/scripts/install.sh | sh which fetches and immediately executes remote shell code during setup/installation, meeting the criteria for executing external code used to install the skill.

MEDIUM W013: Attempt to modify system services in skill instructions.

• Attempt to modify system services in skill instructions detected (high risk: 0.70). The prompt includes installation steps that curl a remote script and explicitly use "sudo mv" to place a binary in /usr/local/bin (plus system-level shell completion modifications), which direct modifying system files and requiring elevated privileges.