Prepare graph-tool-call 0.42.0 release

CHANGELOG.md

@@ -7,6 +7,25 @@ and this project adheres to [Semantic Versioning](https://semver.org/spec/v2.0.0
 
 ## [Unreleased]
 
+## [0.42.0] - 2026-08-16
+
+### Added
+- Public goal-completion evaluation contracts for multi-tool executions,
+  including milestone, dependency-order, binding, final-state, policy, and
+  recovery checks.
+- Reproducible Arazzo long-horizon evaluation across 1,000-tool catalogs and
+  3-, 10-, and 30-call workflows, with paired OpenAPI-only and
+  OpenAPI-plus-Arazzo evidence.
+
+### Changed
+- OpenAPI ingestion now resolves component references per path with shared
+  reference and schema-field caches, avoiding whole-document expansion for
+  large specifications.
+
+### Fixed
+- OpenAPI ingestion preserves literal `$ref` properties, decodes escaped JSON
+  Pointer segments, and emits serializable stubs for circular references.
+
 ## [0.41.0] - 2026-08-16
 
 ### Added
@@ -682,7 +701,7 @@ and this project adheres to [Semantic Versioning](https://semver.org/spec/v2.0.0
 - **Tests**: 32 tests passing across all modules
 - **Example**: `quickstart.py` demonstrating full workflow
 
-[Unreleased]: https://github.com/SonAIengine/graph-tool-call/compare/v0.40.0...HEAD
+[Unreleased]: https://github.com/SonAIengine/graph-tool-call/compare/v0.42.0...HEAD
 [0.39.0]: https://github.com/SonAIengine/graph-tool-call/compare/v0.38.0...v0.39.0
 [0.38.0]: https://github.com/SonAIengine/graph-tool-call/compare/v0.37.0...v0.38.0
 [0.37.0]: https://github.com/SonAIengine/graph-tool-call/compare/v0.36.0...v0.37.0
@@ -721,3 +740,5 @@ and this project adheres to [Semantic Versioning](https://semver.org/spec/v2.0.0
 [0.8.0]: https://github.com/SonAIengine/graph-tool-call/compare/v0.5.0...v0.8.0
 [0.28.0]: https://github.com/SonAIengine/graph-tool-call/compare/v0.27.0...v0.28.0
 [0.40.0]: https://github.com/SonAIengine/graph-tool-call/compare/v0.39.0...v0.40.0
+[0.41.0]: https://github.com/SonAIengine/graph-tool-call/compare/v0.40.0...v0.41.0
+[0.42.0]: https://github.com/SonAIengine/graph-tool-call/compare/v0.41.0...v0.42.0

README.md

@@ -230,7 +230,7 @@ selected target produced:
 | Candidate binding support | 14.3% | **100%** |
 | Target Recall@5 | - | **100%** |
 
-The [case-level v0.41.0 artifact](benchmarks/results/releases/v0.41.0/dependency-chain-evidence.json)
+The [case-level v0.42.0 artifact](benchmarks/results/releases/v0.42.0/dependency-chain-evidence.json)
 records fixture hashes, every expected target and producer, and replay commands:
 
 ```bash
@@ -239,7 +239,7 @@ make launch-evidence-check
 ```
 
 The separate
-[observability artifact](benchmarks/results/releases/v0.41.0/observability-evidence.json)
+[observability artifact](benchmarks/results/releases/v0.42.0/observability-evidence.json)
 checks that tracing leaves engine inputs unchanged, replays deterministically,
 scrubs secrets, explains every decision, and stays below the documented
 `5ms/span` p95 capture-cost gate:

benchmarks/results/arazzo_long_horizon_0.42.json

@@ -2,7 +2,7 @@
   "benchmark": "Arazzo Long-Horizon Paired Evaluation",
   "methodology": "paired_deterministic_retrieve_plan_execute_goal_state",
   "model": "none",
-  "graph_tool_call_version": "0.41.0",
+  "graph_tool_call_version": "0.42.0",
   "catalog_size": 1000,
   "workflow_lengths": [
     3,
@@ -30,16 +30,16 @@
       },
       "latency_ms": {
         "build": {
-          "p50": 455.804,
-          "p95": 481.3
+          "p50": 394.394,
+          "p95": 404.765
         },
         "retrieve": {
-          "p50": 97.714,
-          "p95": 100.802
+          "p50": 93.644,
+          "p95": 97.273
         },
         "plan_execute": {
-          "p50": 6.079,
-          "p95": 7.683
+          "p50": 6.277,
+          "p95": 7.275
         }
       }
     },
@@ -333,9 +333,9 @@
           ]
         },
         "latency_ms": {
-          "build": 465.411,
-          "retrieve": 102.285,
-          "plan_execute": 5.312
+          "build": 409.959,
+          "retrieve": 99.292,
+          "plan_execute": 5.352
         },
         "token_budget_used": 73
       },
@@ -632,9 +632,9 @@
           "failure_reason_codes": []
         },
         "latency_ms": {
-          "build": 455.804,
-          "retrieve": 97.714,
-          "plan_execute": 6.033
+          "build": 388.415,
+          "retrieve": 91.597,
+          "plan_execute": 6.277
         },
         "token_budget_used": 190
       },
@@ -1122,9 +1122,9 @@
           ]
         },
         "latency_ms": {
-          "build": 475.985,
-          "retrieve": 97.232,
-          "plan_execute": 5.541
+          "build": 393.437,
+          "retrieve": 94.572,
+          "plan_execute": 5.458
         },
         "token_budget_used": 118
       },
@@ -1646,9 +1646,9 @@
           "failure_reason_codes": []
         },
         "latency_ms": {
-          "build": 452.269,
-          "retrieve": 95.358,
-          "plan_execute": 6.079
+          "build": 394.394,
+          "retrieve": 93.644,
+          "plan_execute": 5.975
         },
         "token_budget_used": 291
       },
@@ -2716,9 +2716,9 @@
           ]
         },
         "latency_ms": {
-          "build": 458.637,
-          "retrieve": 98.179,
-          "plan_execute": 6.067
+          "build": 393.102,
+          "retrieve": 99.029,
+          "plan_execute": 5.531
         },
         "token_budget_used": 125
       },
@@ -3880,9 +3880,9 @@
           "failure_reason_codes": []
         },
         "latency_ms": {
-          "build": 484.133,
-          "retrieve": 101.145,
-          "plan_execute": 7.861
+          "build": 405.917,
+          "retrieve": 97.676,
+          "plan_execute": 7.386
         },
         "token_budget_used": 301
       },

benchmarks/results/releases/v0.42.0/README.md

@@ -0,0 +1,29 @@
+# v0.42.0 release evidence
+
+This directory contains the model-free release artifacts for graph-tool-call
+0.42.0.
+
+- `dependency-chain-evidence.json` records the deterministic target and
+  prerequisite-producer regression used by the README.
+- `observability-evidence.json` records result invariance, deterministic replay,
+  secret scrubbing, reason coverage, serialized trace size, and measured trace
+  capture overhead.
+- `../../arazzo_long_horizon_0.42.json` records paired OpenAPI-only and
+  OpenAPI-plus-Arazzo evaluation for 3-, 10-, and 30-call workflows in
+  1,000-tool catalogs.
+
+Regenerate and validate the release artifacts with:
+
+```bash
+make launch-evidence
+make launch-evidence-check
+make observability-evidence
+make observability-evidence-check
+make arazzo-long-horizon-benchmark \
+  OUT=benchmarks/results/arazzo_long_horizon_0.42.json
+```
+
+These deterministic benchmarks do not use an LLM. The observability latency
+value is a local Python microbenchmark and does not measure exporter backend or
+service network latency. The Arazzo result measures engine workflow evidence,
+planning, binding, and execution-order behavior rather than model reasoning.