v5.1.8

Runtime Self-Learning v5.1.8

Clean patch release on top of v5.1.7. This does not overwrite the existing v5.1.7 release asset.

Changed

• Adds structured usage evidence for future usage:large_context:* and usage:failed_request:* patterns.
• Appends/deduplicates evidence when existing usage patterns are reinforced.
• Stops self_learning_doctor from counting rejected high-score patterns in evidence_missing diagnostics.
• Synchronizes the release test baseline to 950 tests.

Validation

• npm run build: passed
• npm run check: passed
• npm test: 950 tests, 945 passed, 5 skipped, 0 failed
• npm run benchmark: 17/17 scenarios passed
• npm run perf -- --json: passed, no threshold breaches
• npm run release:check: ready / Score 100
• npm audit: 0 vulnerabilities

Asset

• hanako-runtime-learner-dist.zip
• SHA256: 6490c5879fc1414dc39f4e7dda35355ce9f63c17fc38e88f12f0a40256edcf2f