Inspect and adjust agent runtime parameters like model, iterations, and context window.
npx skills add https://github.com/timothyckl/nano-hermes --skill my
Please help me install this Agent Skill. Skill: my Source: https://github.com/timothyckl/nano-hermes/tree/main/nano_hermes/skills/my Command: npx skills add https://github.com/timothyckl/nano-hermes --skill my
The agent often operates with dynamic runtime settings (model, iterations, context window, token usage, and web configuration). This skill exposes, audits, and adjusts those settings to diagnose issues, optimize performance, and maintain user preferences across turns.
Query the agent's runtime state and adjust model, iterations, context window, or web configuration as needed to resolve issues.
High-intent search queries and answers about installing and using this skill.
To diagnose an agent's runtime state during unexpected pauses, query the current model, iterations, context window, and token usage. The tool reads current settings and scratchpad notes to identify the configuration causing the interruption.
Yes, you can adjust the context window and token usage limits during a task. The tool exposes runtime parameters for modification, applying guardrails to validate parameter ranges and restricting changes to in-memory modifications only.
The best way to check what model an agent is running is to query its runtime state. This tool inspects current configuration settings, including the active model, token usage, and web configuration, reporting the exact parameters.
To tune agent iterations and optimize task performance, inspect the current runtime state and adjust the iteration parameters. The tool validates parameter ranges against guardrails to ensure safe in-memory modifications.
Runtime configuration adjustments remain in-memory only and do not persist permanently. However, the tool can remember user preferences across turns, applying those stored preferences to future responses within the active session.
pradeepmouliExecute a script to generate baseline-spec.md and current-state.md from JSON output.
ScientiaCapitalExport fine-tuned models to GGUF and deploy with vLLM or Ollama.
scawfulDiscover AFS context roots, validate scaffolding, and read or write scratchpad memory.
ibalasiteDetect deployment platforms, infer health checks, and write Deploy Configuration to CLAUDE.md.
