The Agent Skills Directory

[EXTERNAL_DOWNLOADS]: The skill references community Telegram channels using bit.ly URL shorteners. These links are hardcoded and point to established project-related communication channels for the 'it-autonomos' project.
[PROMPT_INJECTION]: The skill ingests content from local files using the Read tool. This presents a surface for indirect prompt injection if source files contain malicious instructions. However, the risk is mitigated by explicit instructions requiring 'STRICT Content Fidelity' and forbidding the agent from adding, removing, or 'improving' content, which limits the likelihood of executing instructions embedded in the source text.
Ingestion points: Reads files from _includes/ and language-specific directories (e.g., ru/).
Boundary markers: None present in the prompt instructions.
Capability inventory: Uses Read, Glob, and Grep tools; instructed to output updated file content for index.md and Jekyll includes.
Sanitization: No automated sanitization; relies on LLM adherence to strict translation rules.

content-translate-article