claude-orchestrator-starter

Orchestrates automatic and manual output quality checks. Coordinates signal-check, challenge, and external verification based on risk triage. Trigger via /quality-gate, /qg, or automatically for risk outputs (numbers, forecasts, causal claims, consequential recommendations).

Captures ideas, notes, links, and thoughts quickly. Use when the user says 'note', 'idea', 'capture', 'save this', 'remember', or wants to save something for later.

Adversarial stress-testing of ideas, plans, decisions, and strategies. Finds weak points, simulates stakeholder objections, surfaces hidden assumptions, and runs pre-mortems. Use when the user says 'challenge', 'stress-test', 'red team', 'objections', 'counter-arguments', 'what could go wrong', 'weaknesses', 'devil's advocate', 'blind spots', 'find holes', or wants to pressure-test thinking before committing.

Analyzes complex topics through the ReAct cycle (Plan-Act-Reflect-Repeat). Structures multi-step investigations with transparent reasoning. Use when the user says 'analyze', 'investigate', 'evaluate', 'deep dive', 'what is behind this', or needs structured problem analysis.

Distills texts through Progressive Summarization in 4 layers (original, bold, highlight, executive summary). Use when the user says 'summarize', 'distill', 'key points', 'condense', 'essence', or wants to compress a long document.

Generates polished outputs from knowledge — articles, posts, presentations, emails. Uses POPP method and UX principles. Use when the user says 'write', 'draft', 'create content', 'blog post', 'formulate', or wants a finished text output.

Saves current session state to a project file for seamless continuation in the next session. Captures decisions, open questions, active context, and next steps. Use when the user says 'handoff', 'save state', 'save session', 'for next time', 'where were we', or at the end of a work session.

Evaluates content substance through 4-axis analysis (meaning, facts, framing, context). Separates signal from noise. Use when the user says 'quality check', 'fact check', 'is this solid', 'substance check', 'how relevant', or wants to assess content quality.