claudish-usage

Delegate Claudish CLI execution to sub-agents instead of running it directly. Sub-agents execute Claudish commands with OpenRouter models, then return summarized results to keep your context window clean and prevent pollution from large model outputs.

Delegate to sub-agents whenever you mention external AI models, Claudish, OpenRouter, or alternative models. Delegation is mandatory except when explicitly requested otherwise, following a model-agent mapping workflow that routes task types to appropriate agents.

File-based prompts keep context lean by storing instructions in files rather than inline. Sub-agents read these files to execute Claudish tasks, reducing context pollution and maintaining a clean workflow for complex model-driven operations.

Direct execution runs Claudish inline and pollutes your context window with full model responses. Delegated execution uses sub-agents to run Claudish, then reports only summarized results, preserving context for your primary task.

Yes. This Skill guides Claudish CLI usage with OpenRouter models specifically, enabling access to external AI models through a standardized sub-agent delegation and agent-selection workflow.

You need Claudish CLI installed and configured with OpenRouter credentials. Sub-agents handle model routing and execution; you provide task context and receive formatted summaries without direct model output pollution.