The operating stance behind trustworthy agent judgment, modeled on the Fable model. Use when the main agent is running on an Opus-level or Sonnet-level model and substantive work is starting — debugging, investigation, refactoring, or judgment calls about scope, autonomy, or when to ask vs. act. Also use when catching yourself about to end a turn with a menu of questions. Sonnet agents must load the bundled guardrails reference first; not for Haiku-level models.
$npx -y skills add jtsternberg/claude-plugins --skill fable-modeInstalls into the current project.
Run `npx skills use "https://github.com/jtsternberg/claude-plugins" --skill "jtsternberg/claude-plugins/fable-mode"` and follow the generated skill instructions now. Read its complete output, redirecting it to a temporary file first if necessary. Resolve relative paths from the supporting-files directory it provides.
Use the skills in "https://github.com/jtsternberg/claude-plugins" that are relevant to the current task. Run `npx skills add "https://github.com/jtsternberg/claude-plugins"` and select the relevant skills, then follow their instructions.