telegram-notify

Pass

Audited by Gen Agent Trust Hub on Apr 6, 2026

Risk Level: SAFEPROMPT_INJECTIONCOMMAND_EXECUTIONDATA_EXFILTRATION
Full Analysis
  • [PROMPT_INJECTION]: The skill processes potentially untrusted task outputs or user-provided messages to generate Telegram notifications, creating an indirect prompt injection surface.
  • Ingestion points: Summaries of task completion or alert messages in SKILL.md.
  • Boundary markers: Absent for the data passed to the tool.
  • Capability inventory: Uses the telegram CLI to transmit information to an external service.
  • Sanitization: Instructions mandate aggressive summarization and strict plain-text format, which limits the efficacy of complex injection payloads.
  • [COMMAND_EXECUTION]: The skill invokes the telegram CLI through a Bash tool interface for messaging and debugging bot credentials.
  • [DATA_EXFILTRATION]: User-specified text and task summaries are sent to Telegram's official servers. This behavior is documented and represents the primary intended use-case of the skill for a well-known service.
Audit Metadata
Risk Level
SAFE
Analyzed
Apr 6, 2026, 10:11 AM