The Agent Skills Directory

[PROMPT_INJECTION]: The skill processes untrusted external data (Telegram messages) to generate AI draft replies, creating an indirect prompt injection surface. This allows content from external messages to potentially influence the agent's behavior. Ingestion points: Telegram messages fetched via scripts/telegram-gramjs.ts (referenced in workflows/process-messages.md). Boundary markers: No explicit delimiters or instructions to ignore embedded commands are specified in the draft generation phase. Capability inventory: The skill can execute local scripts via bun, read/write files in ~/CybosVault/, and perform Telegram API operations. Sanitization: No sanitization of message content before LLM processing is documented.- [COMMAND_EXECUTION]: The skill relies on the execution of local TypeScript scripts using the bun runtime to interact with the Telegram API and manage the local filesystem. Evidence: Multiple commands such as bun scripts/telegram-gramjs.ts and bun scripts/telegram-save-drafts.ts are integrated into the workflows.

telegram