The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill handles data from the Telegram API that can be used to influence the agent's behavior or poison downstream AI integrations.
Ingestion points: Untrusted data enters the system through Telegram Update objects, including user text, callback data, and web app data, as processed in assets/boilerplate/python/bot.py, assets/boilerplate/nodejs/src/handlers.ts, and the AI integration example in references/detailed-guide.md.
Boundary markers: The AI interaction pattern shown in references/detailed-guide.md lacks delimiters or specific instructions to the model to ignore commands embedded within the user-provided text variable.
Capability inventory: The skill includes boilerplate code and utility scripts (scripts/send_message.py, scripts/test_bot.py) that perform network requests to the Telegram Bot API (api.telegram.org). The scripts/setup_project.py script possesses file system write capabilities to initialize project structures.
Sanitization: While the skill correctly escapes HTML for Telegram output, it does not implement input filtering or validation for incoming data before it is used in logic or forwarded to AI adapters.

telegram