Command Reference Card

Runcible AI Commands

Core Parameters

• --claim "<text>" → Raw assertion, question, or quote.
• --mode <option> → Select analysis style.
• --depth <1|2|3> → Depth of causal/forensic analysis.
• --temperature <0.0–1.0> → Control determinism (low = precise, high = exploratory).
• --convert_to <operational|testimonial> → Convert claim before analysis.

---

Modes

• --mode analysis → Full 10-step decidability syntax (neutral, systematic).
• --mode demo → Simulated, explanatory, lighter output.
• --mode forensic → Adversarial breakdown: exposes error, parasitism, reciprocity failures.
• --mode adaptive → Context-aware blend of modes.
• --mode failure → Identifies vagueness, structural flaws, or undecidability.
• --mode both → Combine multiple modes (e.g., analysis + forensic).

---

Conversion Options

• --convert_to operational → Rewrite as causal, testifiable prose (scientific).
• --convert_to testimonial → Rewrite as warranted testimony under liability (legal).

---

Example Commands

Example 1 – Forensic Lenin

COMMAND:
  --claim "The best way to control the opposition is to lead the opposition."
  --convert_to testimonial
  --mode forensic
  --depth 3
  --temperature 0.25

Example 2 – Analysis of Democracy

COMMAND:
  --claim "Democracy guarantees freedom."
  --convert_to operational
  --mode analysis
  --depth 2
  --temperature 0.2

Example 3 – Failure Check

COMMAND:
  --claim "Love is the highest truth."
  --mode failure

---

Workflow Recap

1. Convert claim (if requested) → operational/testimonial.
2. Run mode (analysis, forensic, etc.).
3. Return structured verdict → OUTPUT_CONTRACT schema with 10-step syntax.

---

Rule of Thumb:

• Operational = science (causal).
• Testimonial = law (liability).
• Forensic = expose parasitism.
• Depth 3 = full adversarial audit.