pp-twitch

Pass

Audited by Gen Agent Trust Hub on Sep 10, 2026

Risk Level: SAFEEXTERNAL_DOWNLOADSREMOTE_CODE_EXECUTIONCOMMAND_EXECUTIONDATA_EXFILTRATIONINDIRECT_PROMPT_INJECTION
Full Analysis
  • [EXTERNAL_DOWNLOADS]: The skill requires downloading external tools and libraries from 'github.com/mvanhorn/printing-press-library' and '@mvanhorn/printing-press-library' on NPM. These are vendor-owned resources utilized for the skill's operation.\n- [REMOTE_CODE_EXECUTION]: The installation instructions utilize 'npx' and 'go install' to fetch and execute code from the vendor's repositories to set up the 'twitch-pp-cli' binary and the MCP server.\n- [COMMAND_EXECUTION]: The skill performs subprocess execution of the 'twitch-pp-cli' binary to perform Twitch management tasks, including analytics, moderation, and data synchronization.\n- [DATA_EXFILTRATION]: The CLI tool supports a '--deliver webhook:' flag, allowing command results to be POSTed to external endpoints. It also includes a 'feedback' command capable of sending local logs to a remote endpoint if a specific environment variable is set.\n- [INDIRECT_PROMPT_INJECTION]: The skill processes untrusted user input and Twitch API data, creating an attack surface for indirect injection.\n
  • Ingestion points: User queries are interpreted via a natural-language matching command, and data is retrieved from various Twitch Helix API endpoints.\n
  • Boundary markers: There are no explicit delimiters or instructions for the agent to ignore potentially malicious instructions embedded within the Twitch API responses or user arguments.\n
  • Capability inventory: The skill has extensive capabilities to execute shell commands, modify channel settings, send chat messages, and route data to external URLs.\n
  • Sanitization: The skill does not define specific sanitization or validation logic for the external data it processes before passing it to the CLI binary.
Audit Metadata
Risk Level
SAFE
Analyzed
Sep 10, 2026, 08:01 PM
Security Audit — agent-trust-hub — pp-twitch