judge

sigmap judge tells you whether an answer appears to be supported by the context you actually supplied.

sigmap judge --response response.txt --context .context/query-context.md
sigmap judge --response response.txt --context .context/query-context.md --json
sigmap judge --response response.txt --context .context/query-context.md --learn

What it reports

• Groundedness — overlap score between the answer and the source context
• Support level — high / medium / low based on groundedness ratio
• Unsupported symbols — tokens and claims that look weakly supported

This is a traceability check, not a truth oracle. It helps answer: "Did this response come from the code I provided, or did the model drift?"

Typical output

Groundedness       : 78%
Support level      : HIGH
Unsupported symbols: none

Complete workflow: ask → get answer → judge

Step 1: Generate focused context

sigmap ask "explain the auth flow"
# Creates: .context/query-context.md

Step 2: Get AI response

# Copy .context/query-context.md into your AI chat
cat .context/query-context.md
# Paste the output into Claude, Copilot, ChatGPT, or your IDE
# Ask your question: "Explain the auth flow"
# Copy the AI's response

Step 3: Save the response

# Create response.txt with the AI's answer
cat > response.txt << 'EOF'
[Paste the complete AI response here]
EOF

Step 4: Judge groundedness

sigmap judge --response response.txt --context .context/query-context.md
# Output: Groundedness score and support level

Opt-in learning

With --learn, judge can apply a small local boost or penalty to the files referenced in the context headings:

• strongly grounded result → small boost
• weakly grounded result → small penalty
• middle band → no change

This learning is local-only and stored in .context/weights.json.

When to use it

• reviewing AI-generated explanations
• checking whether a debugging suggestion is really grounded in the shown files
• grading prompt/response pairs in demos or release benchmarks
• feeding the learning engine carefully instead of manually every time