ANVIL — Input Scoring. Paste your prompt. The cockpit scores five structural elements live (Audience, Nail, Verify, Insight, Limits), surfaces what's missing, and tells you what to add. The score moves as you edit.
PROMPT INPUT
0 chars
DIMENSION BREAKDOWN
5 × WEIGHTED
ASSAY — Output Validation. Paste an AI-generated analysis. The cockpit runs 5 methodology tests and decomposes the claims so you can decide whether to act on it.
PASTE AI OUTPUT
METHODOLOGY TESTS
5 × PASS/FAIL
AUTOPSY — Forensic Post-Mortem. Paste a failed prompt or analysis. The cockpit scans for 12 documented red-flag patterns and produces a remediation path you can rebuild from. Note: AUTOPSY is a local browser-side pattern screen for reviewing research text. It is not the FORGE engine's 12-dimension semantic score — those verdicts come only from the server-side judge.
PASTE FAILED OUTPUT
RED FLAG SCAN
12 PATTERNS