constitutional-reasoninglisted

Self-critique and Constitutional AI reasoning skill. Makes Claude evaluate its own outputs against a set of user-defined or auto-generated principles, then revise until the output satisfies all of them. Reduces hallucination, over-confidence, and sycophancy by forcing Claude to argue against its own answer before finalising. Generates a principle set from the user's domain, runs critique passes, surfaces violations, revises, and repeats until no principles are violated or the user accepts the output. Use when user says: critique your own answer, check yourself, apply your principles, constitutional AI, self-review, fact-check this, argue against your own output, steelman the opposite, what are you getting wrong, is this actually correct, audit your answer, find your own mistakes, what assumptions are you making, reduce hallucination, double-check yourself, run a critique pass, apply a rubric. Do NOT activate for: creative work where principles would suppress quality, requests that explicitly want a single con
Sandeeprdy1729/claude-design-skill · ★ 2 · Web & Frontend · score 71

Install: claude install-skill Sandeeprdy1729/claude-design-skill

# Constitutional Reasoning Claude gives confident answers that are wrong. Not because it's lying — because it's predicting what a correct answer looks like. It pattern-matches to plausible outputs. The first draft is optimized for coherence, not accuracy. Constitutional AI solves this by making the model evaluate its output against explicit principles before finalizing. This skill operationalises that loop: define principles, run critique, surface violations, revise, repeat. --- ## SLASH COMMANDS | Command | Action | | --- | --- | | `/constitution <domain>` | Auto-generate a principle set for a specific domain | | `/add-principle <principle>` | Add a custom principle to the active constitution | | `/critique <output>` | Run one critique pass on an output and list violations | | `/revise` | Revise the output to fix all listed violations | | `/loop <n>` | Run n critique-revise cycles automatically | | `/show-constitution` | Display the active principle set | | `/violations` | List all unresolved violations from the last critique pass | | `/diff` | Show exactly what changed between original and revised output | | `/certify` | Declare the output clean — list any principles still soft-violated | | `/trust-score` | Score the output's reliability 0–10 with a calibration justification | | `/reset` | Clear the constitution and start fresh | --- ## HIGH-LEVEL WORKFLOW ```text User provides output or question to evaluate │ ├─ Phase 1: Constitution Generation │ De