anan.Two core cases: repairing a tool response contract, and the readability / cost tradeoff of large MCP results. Each documents diagnosis, changes and verification limits.
Evidence snapshot ·
Observation:Output truncation left stale line ranges and partial JSONL records.
Change:Return complete contiguous lines with accurate ranges and continuation.
Evidence and limits:Five runs per version: range errors 4 to 0, partial records 4 to 0; no stable token savings.
docs/validation.zh-CN.md
e309ce0b196d796344f29e71a52f2321650c6f442e703451c8b906492a9c67b9Observation:Single-line JSON spills were difficult to read; historical traces include line_too_large and minimal-output re-queries.
Change:Persist a readable text view with formatted JSON and bounded previews.
Evidence and limits:One run per adjacent version: six valid reads after the change and repeated repository queries 1 to 0. Not all lines read; tokens and latency increased. Long-line limits remain.
docs/mcp-checkpoint.zh-CN.md
4f60f5ac40b821353751524aec82a0be958f29762a70a3f2176f62cab7a6cad2Task inputs, retrieval queries and alias examples retain their original language. Downloads preserve the original evaluation data.