The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill is designed to ingest and analyze untrusted data, specifically user intents and the descriptions of other skills. It contains specific instructions to capture intent as a single sentence and match it against trigger branches. While this creates a surface for indirect prompt injection, the skill explicitly mandates a 'Route, don't do' policy, stating it never performs the work itself and only identifies the correct tool. This functional limitation significantly mitigates the risk of an injection influencing dangerous downstream actions through this specific skill.
[NO_CODE]: The skill consists entirely of instructional Markdown and YAML metadata. It does not include any scripts (Python, JavaScript, shell), executable binaries, or configuration files that trigger command execution. All routing logic is performed by the LLM based on the provided instructions.

nextreme-router