skills/mikeyobrien/rho/tasker-xml/Gen Agent Trust Hub

tasker-xml

Pass

Audited by Gen Agent Trust Hub on Aug 30, 2026

Risk Level: SAFECOMMAND_EXECUTIONPRIVILEGE_ESCALATIONDATA_EXFILTRATIONINDIRECT_PROMPT_INJECTION
Full Analysis
  • [COMMAND_EXECUTION]: The skill provides detailed templates for the Tasker 'Run Shell' action (code 123), enabling the generation of automation that executes arbitrary shell commands on an Android device.
  • [PRIVILEGE_ESCALATION]: The documentation specifically includes instructions for the 'use root' parameter within shell command actions, allowing for the creation of automation that operates with administrative privileges.
  • [DATA_EXFILTRATION]: Templates for 'Browse URL' (code 104) and 'Write File' (code 410) are included, which provide mechanisms to send data to external URLs or write sensitive information to the device's storage (e.g., /storage/emulated/0/Download/).
  • [INDIRECT_PROMPT_INJECTION]:
  • Ingestion points: The skill documents how to process untrusted data from Android Intents (Intent Received event code 599) and screen text content via the AutoInput plugin (UI Query action).
  • Boundary markers: The provided XML templates do not include delimiters or instructions to treat data from %aitext() or Intent extras as untrusted content.
  • Capability inventory: The skill provides access to shell commands (including root), file system writes, and web navigation.
  • Sanitization: There are no documented steps for sanitizing or validating external input before it is used in subsequent automation steps or written to result files for the agent to process.
Audit Metadata
Risk Level
SAFE
Analyzed
Aug 30, 2026, 11:57 AM
Security Audit — agent-trust-hub — tasker-xml