bot-channel-task

Pass

Audited by Gen Agent Trust Hub on Apr 10, 2026

Risk Level: SAFEPROMPT_INJECTIONCOMMAND_EXECUTION
Full Analysis
  • [PROMPT_INJECTION]: Indirect prompt injection surface identified where external input could influence agent behavior.
  • Ingestion points: Processes data received via external messaging channels (e.g., Telegram) as described in SKILL.md.
  • Boundary markers: No explicit delimiters or instructions to ignore embedded commands within the ingested channel data are provided.
  • Capability inventory: The skill can launch background subagents using the Agent tool, execute internal job management commands (/bot-jobs), and send replies to the messaging channel.
  • Sanitization: No evidence of sanitization, validation, or filtering of the external message content before processing.
  • [COMMAND_EXECUTION]: The skill interprets natural language user intent to invoke internal system subcommands (/bot-jobs add, /bot-jobs list, /bot-jobs pause) for managing scheduled tasks.
Audit Metadata
Risk Level
SAFE
Analyzed
Apr 10, 2026, 07:55 PM
Security Audit — agent-trust-hub — bot-channel-task