telegram-serverless-bot
Pass
Audited by Gen Agent Trust Hub on Oct 15, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill creates an attack surface by ingesting and processing untrusted data from the Telegram Bot API, including message text, callback data, and media captions.
- Ingestion points: Untrusted data enters the agent's context through the Telegram update handlers defined in
tgcloud/handlers/message.jsandtgcloud/handlers/callback_query.js. - Boundary markers: The skill documentation explicitly instructs developers to use parameterized query builders for database operations and to escape HTML/Markdown entities in outgoing messages to maintain data boundaries.
- Capability inventory: The skill utilizes a built-in SQLite database (
sdk/db) for state management, a Telegram API client (sdk/api), and an outbound HTTP fetch client (sdk/fetch). - Sanitization: Practical sanitization techniques are provided, including an
escapeHtmlfunction for display data and the enforcement of the platform's query builder to prevent SQL injection.
Audit Metadata