QueryReceipts 0.2.1

Sanitization patch on 0.2.0: fixture identifiers neutralized to the synthetic FleetDB vocabulary following a Sonnet-based sensitivity review (capture-derived names, runner labels, CU build numbers). No functional changes; 58 unit + 4 integration tests green locally and in CI.

QueryReceipts 0.2.0

The pre-approved roadmap, shipped: driver transport (receipts run — executes prescriptions via your runner command, secrets never enter the ledger), MCP transport (receipts mcp-serve — the full toolset as a stdio MCP server), the PostgreSQL engine pack (EXPLAIN ANALYZE JSON parsing with loops-aware skew, CTE-native validation, proven by a second full-loop e2e), and CI mode (receipts verify tamper gate + this repo's own CI re-proving both engines' cure loops on every push). 58 unit + 4 integration tests.

QueryReceipts 0.1.0

Initial release: the verified-fix cure loop for SQL Server, proven end to end against a live instance. Equivalence proofs with comparability gates (CTE-capable), pre-pinned benchmark protocols, three-valued certificates citing every artifact hash, parser-first evidence handling, and a Claude Code plugin skin. See docs/examples/fleetdb-walkthrough.md for a real certificate: reads −99.8%, elapsed −91.3%, 16/16 equivalence checks.