search-messages
Pass
Audited by Gen Agent Trust Hub on Nov 7, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONDYNAMIC_EXECUTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill processes Telegram messages, which are untrusted external data sources that could contain malicious instructions designed to influence the agent. • Ingestion points: Telegram messages fetched from external chats via the tg-cli.ts script in SKILL.md. • Boundary markers: The output is formatted with specific headers, but message bodies are included without specific delimiters or warnings to ignore embedded instructions. • Capability inventory: The skill has access to the Bash tool to execute commands. • Sanitization: No sanitization of the retrieved message content is mentioned or performed before it is presented to the agent.
- [DYNAMIC_EXECUTION]: The skill instructs the agent to self-evolve by modifying the SKILL.md file directly if issues are encountered. This self-modification of the skill's own source code at runtime is a risk factor as it allows persistent changes in the agent's logic based on execution context.
Audit Metadata