skills/qwibitai/nanoclaw/add-signal/Gen Agent Trust Hub

add-signal

Pass

Audited by Gen Agent Trust Hub on Apr 26, 2026

Risk Level: SAFECOMMAND_EXECUTIONEXTERNAL_DOWNLOADSREMOTE_CODE_EXECUTIONPROMPT_INJECTION
Full Analysis
  • [COMMAND_EXECUTION]: Uses sudo with system package managers (apt-get, dnf) to install Java 17, which is a required dependency for the Signal CLI.
  • [COMMAND_EXECUTION]: Manages the application daemon using platform-specific service managers (launchctl on macOS and systemctl on Linux) to ensure the Signal integration runs as a background process.
  • [EXTERNAL_DOWNLOADS]: Fetches the signal-cli binary and associated version metadata from GitHub's official API and releases pages.
  • [REMOTE_CODE_EXECUTION]: Executes a Python snippet via the command line to parse version information from a GitHub API response. While this involves piping data to an interpreter, it is a localized operation for configuration purposes rather than the execution of remote scripts.
  • [PROMPT_INJECTION]: The skill establishes a communication channel that ingests data from Signal messages, creating a surface for indirect prompt injection.
  • Ingestion points: Incoming messages from Signal users are received via the signal-cli daemon and processed by the system (referenced in src/channels/signal.ts).
  • Boundary markers: The setup instructions do not detail the use of specific delimiters or markers to separate external user input from system instructions.
  • Capability inventory: The environment allows for file system writes to project source directories, shell command execution via binary management, and local database modifications via sqlite3.
  • Sanitization: No specific input sanitization or validation steps for incoming messaging data are described in the provided configuration files.
Audit Metadata
Risk Level
SAFE
Analyzed
Apr 26, 2026, 09:50 PM