agent-runtime-kit 0.5.1__tar.gz → 0.5.2__tar.gz

{agent_runtime_kit-0.5.1 → agent_runtime_kit-0.5.2}/CHANGELOG.md

@@ -5,6 +5,32 @@ All notable changes to this project are documented in this file.
 The format is based on [Keep a Changelog](https://keepachangelog.com/en/1.1.0/),
 and this project adheres to [Semantic Versioning](https://semver.org/spec/v2.0.0.html).
 
+## 0.5.2 - 2026-09-19
+
+### Changed
+
+- Updated tested Claude Agent SDK to `0.2.157`, Google Antigravity SDK to
+  `0.1.17`, and the Codex SDK and its exactly coupled CLI to `0.154.0`.
+- SDK evolution supports explicit model, reasoning effort and analysis executable
+  options, preserves older locked baselines in implementation history, and
+  constructs real Claude options and Codex wire parameters in candidate probes.
+- Upgrade skill invocations now include necessary runtime adaptations,
+  verification, and a regular pull request. Explicit report-only and local-only
+  requests retain their narrower scope.
+
+### Fixed
+
+- Antigravity permissive deny-lists preserve all unrelated tools when the SDK's
+  default tool set changes, including the `ask_question` change in 0.1.17.
+- Codex release evidence uses the current official documentation endpoints.
+- Antigravity preserves explicit `vertex=True` with an API key for Vertex
+  Express authentication, rejects mixed regional credentials, and refuses older
+  SDKs or configurations that would silently drop the requested endpoint.
+- Release-note collection recognizes versions present only in a GitHub
+  discussion title and rejects unrelated or longer version-prefix matches.
+- PyPI publication validates and uploads Core Metadata 2.5 distributions with
+  matching Twine support.
+
 ## 0.5.1 - 2026-08-29
 
 ### Added

{agent_runtime_kit-0.5.1 → agent_runtime_kit-0.5.2}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.5
 Name: agent-runtime-kit
-Version: 0.5.1
+Version: 0.5.2
 Summary: One typed runtime API for Claude, Codex, and Antigravity agent SDKs.
 Project-URL: Homepage, https://github.com/ebarti/agent-runtime-kit
 Project-URL: Repository, https://github.com/ebarti/agent-runtime-kit
@@ -22,13 +22,13 @@ Requires-Dist: jsonschema<5,>=4.18
 Provides-Extra: all
 Requires-Dist: claude-agent-sdk<0.3,>=0.2.87; extra == 'all'
 Requires-Dist: google-antigravity<0.2,>=0.1.2; extra == 'all'
-Requires-Dist: openai-codex<0.148,>=0.1.0b3; extra == 'all'
+Requires-Dist: openai-codex<0.155,>=0.1.0b3; extra == 'all'
 Provides-Extra: antigravity
 Requires-Dist: google-antigravity<0.2,>=0.1.2; extra == 'antigravity'
 Provides-Extra: claude
 Requires-Dist: claude-agent-sdk<0.3,>=0.2.87; extra == 'claude'
 Provides-Extra: codex
-Requires-Dist: openai-codex<0.148,>=0.1.0b3; extra == 'codex'
+Requires-Dist: openai-codex<0.155,>=0.1.0b3; extra == 'codex'
 Description-Content-Type: text/markdown
 
 # agent-runtime-kit

{agent_runtime_kit-0.5.1 → agent_runtime_kit-0.5.2}/docs/providers.md

@@ -188,16 +188,32 @@ constructor, `GEMINI_API_KEY`, or `GOOGLE_API_KEY`. Without an API key, the
 adapter uses Vertex AI config with Google ADC when a project can be discovered
 from ADC, `GOOGLE_CLOUD_PROJECT`, or `GCLOUD_PROJECT`; location defaults to
 `global` unless `GOOGLE_CLOUD_LOCATION`, `GOOGLE_CLOUD_REGION`, or
-`CLOUD_ML_REGION` is set. Precedence: an explicit constructor `api_key` is the
-most specific request and wins; otherwise an explicit
+`CLOUD_ML_REGION` is set. An explicit constructor `api_key` takes precedence over
+ambient credentials. With `vertex=True`, that key selects Vertex Express mode
+(supported by released `google-antigravity>=0.1.16` versions): no ADC, project,
+or location is required. Older or unverifiable SDK releases raise
+`UnsupportedTaskInputError` for Express mode; their existing authentication
+modes remain available. Explicit project/location arguments cannot be combined
+with Express mode. For example:
+
+```python
+runtime = AntigravityAgentRuntime(vertex=True, api_key=vertex_express_key)
+```
+
+Without an explicit key, an explicit
 `AntigravityAgentRuntime(vertex=True, ...)` takes precedence over an ambient
 `GEMINI_API_KEY`/`GOOGLE_API_KEY`, so a Vertex-configured runtime is never
-silently redirected to the Gemini API by an exported environment key. Under
-`PERMISSIVE`, `disallowed_tools` maps to `CapabilitiesConfig.disabled_tools`
-(the baseline is every tool, so "enable everything else" is exact); under any
-other posture the deny-list is folded into an allow-list of the mode's baseline
-minus the denied tools, so denying one tool can never re-enable others past the
-baseline. An allow-list and a deny-list are mutually exclusive (the SDK forbids
+silently redirected to the Gemini API by an exported environment key. When no
+model is supplied, the adapter preserves the vendor default (`gemini-3.8-flash`
+in Antigravity 0.1.16); supply `default_model` or `AgentTask.model` when a stable
+model selection is required. Under
+`PERMISSIVE`, `disallowed_tools` becomes an explicit `enabled_tools` list of all
+built-in tools minus the denied tools. Antigravity 0.1.17 changed its implicit
+deny-list baseline to exclude `ask_question`; using an explicit list preserves
+the adapter's existing all-tools contract. Under other postures the deny-list
+is subtracted from the mode's narrower baseline, so denying one tool can never
+re-enable others past that baseline. An allow-list and a deny-list are mutually
+exclusive (the SDK forbids
 combining enabled and disabled tool lists), so supplying both is rejected. Tool
 names are validated against the
 `BuiltinTools` enum (`"view_file"`, not `"Read"`). `budget_usd` and

{agent_runtime_kit-0.5.1 → agent_runtime_kit-0.5.2}/docs/sdk-evolution-agent.md

@@ -4,6 +4,15 @@ The SDK evolution agent is a local dogfood workflow for keeping
 agent-runtime-kit aligned with Claude Agent SDK, OpenAI Codex SDK, and Google
 Antigravity SDK as those upstream packages evolve.
 
+For a complete upgrade, invoke the repository's
+[`agent-runtime-kit-upgrade` skill](../.codex/skills/agent-runtime-kit-upgrade/SKILL.md).
+It collects evidence first, maps upstream implementation trends to the current
+integration, implements the necessary adapter/runtime changes with tests and
+docs, applies the compatible dependency update, and creates a verified regular
+pull request. Explicit report-only and local-only requests retain those limits.
+The CLI examples below describe the underlying runner; its report-only default
+and dependency-only implementation lane are stages within that broader skill.
+
 For the intended architecture, evidence contract, behavior probe strategy,
 changelog strategy, caveats, and alternatives, see
 [`docs/sdk-evolution-agent-design.md`](sdk-evolution-agent-design.md).
@@ -57,13 +66,26 @@ env -u UV_EXCLUDE_NEWER \
   uv run --locked --extra codex python -m examples.sdk_evolution_agent.auth ensure-codex
 ```
 
-Codex-backed SDK evolution runs explicitly choose `gpt-5.5` with
+Codex-backed SDK evolution runs default to `gpt-5.5` with
 `reasoning_effort=xhigh` for the AI stages that analyze direction, decide the
 update plan, and review the result. Dependency application and verification are
 deterministic. This model policy
 is applied only to `codex-agent-sdk`; Claude and Antigravity runs keep their
 provider-native model selection because `gpt-5.5` is not a valid model override
 for those adapters.
+Use `--model` and `--reasoning-effort` to select the requested model explicitly
+for every AI stage. For example, use `--model gpt-6-astra --reasoning-effort xhigh`.
+These options populate the first-class `AgentTask` fields and are recorded in
+`config.json`; unsupported provider values remain provider errors.
+
+If the locked Codex CLI cannot run the requested analysis model, the optional
+`--codex-bin /absolute/path/to/codex` selects an installed executable through
+the SDK's supported `CodexConfig.codex_bin`. It applies only to the analysis
+driver and is recorded separately in `config.json`. Exact baseline/candidate
+inspection and the SDK/CLI dependency coupling still use the published package
+versions. Omit this override for the required post-update run against the final
+locked SDK and bundled CLI. The override requires `--runtime codex-agent-sdk`
+and cannot be combined with an injected runtime or registry.
 Vendor runtimes also enable `reuse_process=True` for the multi-stage run, so
 compatible SDK subprocesses can stay warm across analysis/review stages and are
 closed by the CLI when the internally owned runtime exits. Codex reuses its
@@ -148,7 +170,15 @@ failures are retried. Candidate inspection writes the exact compatibility
 transition to `api_diffs.json` and executes the adapter contract against that
 same version. Independently, recent-release inspection fingerprints shipped
 files and Python AST definitions, compares adjacent releases, and writes the
-result to `implementation_diffs.json`. Source is never copied into the report;
+result to `implementation_diffs.json`. History also includes the exact locked
+baseline and candidate when they fall outside the latest three releases, so
+the changes between the installed baseline and the recent releases remain
+visible. Claude probes construct permission, budget, session, MCP and schema
+options; Codex probes construct configuration and validated thread/turn wire
+parameters, including permissions, reasoning effort and structured output.
+These checks complement real-adapter regression tests and live runtime checks;
+they do not establish behavior inside opaque executables.
+Source is never copied into the report;
 only paths, counts, sizes, and hashes are retained. Candidate inspection is
 opt-in because it executes freshly downloaded upstream code; historical
 inspection uses the same explicit consent. Without the flag, missing trend

{agent_runtime_kit-0.5.1 → agent_runtime_kit-0.5.2}/examples/sdk_evolution_agent/behavior.py

@@ -868,18 +868,26 @@ def _probe_claude(*, version: str | None, scope: str) -> BehaviorProbeResult:
         "output_format",
     }
     missing = sorted(expected - fields)
+    cases, failures = _configuration_contract_cases(package)
+    failed = bool(missing or failures)
     return BehaviorProbeResult(
         package=package,
         version=version,
         scope=scope,
         probe="adapter-contract",
-        status="fail" if missing else "pass",
+        status="fail" if failed else "pass",
         summary=(
-            "ClaudeAgentOptions exposes required adapter fields."
-            if not missing
-            else "ClaudeAgentOptions is missing required adapter fields."
+            "ClaudeAgentOptions exposes and constructs required adapter options."
+            if not failed
+            else "ClaudeAgentOptions rejected the required adapter contract."
         ),
-        details={"fields": sorted(fields), "required_fields": sorted(expected), "missing": missing},
+        details={
+            "fields": sorted(fields),
+            "required_fields": sorted(expected),
+            "missing": missing,
+            "construction_cases": cases,
+            "construction_failures": failures,
+        },
     )
 
 
@@ -896,16 +904,18 @@ def _probe_codex(*, version: str | None, scope: str) -> BehaviorProbeResult:
     missing_run = sorted(expected_run - run_params)
     missing_start = sorted(expected_start - start_params)
     missing = missing_run + [f"thread_start.{item}" for item in missing_start]
+    cases, failures = _configuration_contract_cases(package)
+    failed = bool(missing or failures)
     return BehaviorProbeResult(
         package=package,
         version=version,
         scope=scope,
         probe="adapter-contract",
-        status="fail" if missing else "pass",
+        status="fail" if failed else "pass",
         summary=(
-            "Codex thread APIs expose required adapter parameters."
-            if not missing
-            else "Codex thread APIs are missing required adapter parameters."
+            "Codex configuration and wire parameters construct the required adapter contract."
+            if not failed
+            else "Codex configuration or thread APIs rejected the required adapter contract."
         ),
         details={
             "run_params": sorted(run_params),
@@ -913,10 +923,76 @@ def _probe_codex(*, version: str | None, scope: str) -> BehaviorProbeResult:
             "required_run_params": sorted(expected_run),
             "required_start_params": sorted(expected_start),
             "missing": missing,
+            "construction_cases": cases,
+            "construction_failures": failures,
         },
     )
 
 
+def _configuration_contract_cases(package: str) -> tuple[list[str], list[str]]:
+    """Construct real provider values without starting clients or reading auth.
+
+    This self-contained function is also embedded in isolated candidate probes.
+    """
+    import importlib
+
+    cases: list[str] = []
+    failures: list[str] = []
+    try:
+        if package == "claude-agent-sdk":
+            cases.append("claude-options")
+            sdk = importlib.import_module("claude_agent_sdk")
+            options = sdk.ClaudeAgentOptions(
+                permission_mode="plan",
+                allowed_tools=["Read"],
+                disallowed_tools=["Bash"],
+                system_prompt="Inspect only",
+                cwd=".",
+                resume="test-session",
+                max_budget_usd=1.0,
+                env={},
+                mcp_servers={"files": {"type": "stdio", "command": "test-mcp", "args": []}},
+                output_format={"type": "json_schema", "schema": {"type": "object"}},
+            )
+            if options.permission_mode != "plan" or options.max_budget_usd != 1.0:
+                raise ValueError("permission or budget was not preserved")
+        elif package == "openai-codex":
+            sdk = importlib.import_module("openai_codex")
+            wire = importlib.import_module("openai_codex.generated.v2_all")
+            cases.append("codex-config")
+            config = sdk.CodexConfig(cwd=".", config_overrides=("features.plugins=false",), env={})
+            if config.cwd != ".":
+                raise ValueError("working directory was not preserved")
+            cases.append("codex-thread-permissions")
+            thread = wire.ThreadStartParams(
+                cwd=".",
+                developer_instructions="Inspect only",
+                model="requested-model",
+                approval_policy="never",
+                sandbox="read-only",
+            )
+            if thread.model_dump(mode="json", by_alias=True)["sandbox"] != "read-only":
+                raise ValueError("thread sandbox was not preserved")
+            cases.append("codex-turn-schema-and-effort")
+            turn = wire.TurnStartParams(
+                thread_id="test-thread",
+                input=[],
+                cwd=".",
+                model="requested-model",
+                approval_policy="never",
+                sandbox_policy={"type": "readOnly"},
+                effort="xhigh",
+                output_schema={"type": "object"},
+            )
+            values = turn.model_dump(mode="json", by_alias=True)
+            if values["effort"] != "xhigh" or values["outputSchema"] != {"type": "object"}:
+                raise ValueError("reasoning effort or schema was not preserved")
+    except Exception as exc:
+        # These values are fixed fixtures, but keep diagnostics free of local paths.
+        failures.append(f"{cases[-1] if cases else package}:{type(exc).__name__}")
+    return cases, failures
+
+
 def _probe_codex_cli_bin(*, version: str | None, scope: str) -> BehaviorProbeResult:
     package = "openai-codex-cli-bin"
     module_name = DEFAULT_MODULES[package]
@@ -1206,8 +1282,11 @@ def _string_or_none(value: object) -> str | None:
     return text or None
 
 
-_PROBE_SCRIPT = textwrap.dedent(
-    r"""
+_PROBE_SCRIPT = (
+    inspect.getsource(_configuration_contract_cases)
+    + "\n"
+    + textwrap.dedent(
+        r"""
     import importlib
     import importlib.metadata
     import inspect
@@ -1280,15 +1359,20 @@ _PROBE_SCRIPT = textwrap.dedent(
                 "max_budget_usd", "output_format",
             }
             missing = sorted(expected - option_fields)
+            cases, failures = _configuration_contract_cases(package)
+            contract_failed = bool(missing or failures)
             payload = [result(
                 "adapter-contract",
-                "fail" if missing else "pass",
-                "ClaudeAgentOptions exposes required adapter fields." if not missing
-                else "ClaudeAgentOptions is missing required adapter fields.",
+                "fail" if contract_failed else "pass",
+                "ClaudeAgentOptions exposes and constructs required adapter options."
+                if not contract_failed
+                else "ClaudeAgentOptions rejected the required adapter contract.",
                 {
                     "fields": sorted(option_fields),
                     "required_fields": sorted(expected),
                     "missing": missing,
+                    "construction_cases": cases,
+                    "construction_failures": failures,
                 },
             )]
         elif package == "openai-codex":
@@ -1300,17 +1384,22 @@ _PROBE_SCRIPT = textwrap.dedent(
             missing_run = sorted(expected_run - run_params)
             missing_start = sorted(expected_start - start_params)
             missing = missing_run + [f"thread_start.{item}" for item in missing_start]
+            cases, failures = _configuration_contract_cases(package)
+            contract_failed = bool(missing or failures)
             payload = [result(
                 "adapter-contract",
-                "fail" if missing else "pass",
-                "Codex thread APIs expose required adapter parameters." if not missing
-                else "Codex thread APIs are missing required adapter parameters.",
+                "fail" if contract_failed else "pass",
+                "Codex configuration and wire parameters construct the required adapter contract."
+                if not contract_failed
+                else "Codex configuration or thread APIs rejected the required adapter contract.",
                 {
                     "run_params": sorted(run_params),
                     "start_params": sorted(start_params),
                     "required_run_params": sorted(expected_run),
                     "required_start_params": sorted(expected_start),
                     "missing": missing,
+                    "construction_cases": cases,
+                    "construction_failures": failures,
                 },
             )]
         elif package == "openai-codex-cli-bin":
@@ -1395,4 +1484,5 @@ _PROBE_SCRIPT = textwrap.dedent(
 
     print(json.dumps(payload, sort_keys=True))
     """
-).strip()
+    ).strip()
+)

{agent_runtime_kit-0.5.1 → agent_runtime_kit-0.5.2}/examples/sdk_evolution_agent/cli.py

@@ -86,6 +86,14 @@ def parse_args(argv: list[str] | None = None) -> RunOptions:
         dest="packages",
         help="Vendor SDK package to inspect. Repeat to inspect multiple packages.",
     )
+    parser.add_argument("--model", help="Explicit model for every AI analysis and review stage.")
+    parser.add_argument(
+        "--reasoning-effort", help="Provider-native reasoning effort for AI stages."
+    )
+    parser.add_argument(
+        "--codex-bin",
+        help="Codex analysis executable override; does not change inspected SDK/CLI versions.",
+    )
     parser.add_argument(
         "--report-dir",
         type=Path,
@@ -128,6 +136,8 @@ def parse_args(argv: list[str] | None = None) -> RunOptions:
         help="Draft PR title.",
     )
     args = parser.parse_args(argv)
+    if args.codex_bin and args.runtime != "codex-agent-sdk":
+        parser.error("--codex-bin requires --runtime codex-agent-sdk")
     return RunOptions(
         workspace=Path.cwd(),
         runtime=args.runtime,
@@ -142,6 +152,9 @@ def parse_args(argv: list[str] | None = None) -> RunOptions:
         pr_base=args.pr_base,
         commit_message=args.commit_message,
         pr_title=args.pr_title,
+        model=args.model,
+        reasoning_effort=args.reasoning_effort,
+        codex_bin=args.codex_bin,
     )
 
 
@@ -168,8 +181,14 @@ async def run_agent(
         implementation_enabled=options.implementation_enabled,
         draft_pr=options.draft_pr,
         event_sink=event_sink,
+        model=options.model,
+        reasoning_effort=options.reasoning_effort,
+    )
+    if options.codex_bin is not None and (runtime is not None or registry is not None):
+        raise ValueError("--codex-bin cannot be combined with an injected runtime or registry")
+    selected_runtime = runtime or resolve_runtime(
+        options.runtime, registry=registry, codex_bin=options.codex_bin
     )
-    selected_runtime = runtime or resolve_runtime(options.runtime, registry=registry)
     close_owned_runtime = runtime is None
     try:
         pre_run_results: list[dict[str, Any]] = []
@@ -245,6 +264,8 @@ async def run_agent(
                 implementation_enabled=context.implementation_enabled,
                 draft_pr=context.draft_pr,
                 event_sink=event_sink,
+                model=options.model,
+                reasoning_effort=options.reasoning_effort,
             ),
         )
         implementation = await maybe_run_implementation(
@@ -728,15 +749,15 @@ def _collect_implementation_snapshots(
             ]
             history.extend(matching[:1])
             continue
-        versions = _ordered_versions(recent)
-        if len(versions) < 2:
-            versions = _ordered_versions(
-                [
-                    *recent,
-                    str(package.get("locked_version") or ""),
-                    str(package.get("candidate_version") or ""),
-                ]
-            )
+        # The locked baseline can fall outside the latest three releases. Always
+        # include it so direction analysis covers the actual upgrade interval.
+        versions = _ordered_versions(
+            [
+                *recent,
+                str(package.get("locked_version") or ""),
+                str(package.get("candidate_version") or ""),
+            ]
+        )
         for version in versions:
             snapshot = by_exact_version.get((name, version))
             if snapshot is None:

{agent_runtime_kit-0.5.1 → agent_runtime_kit-0.5.2}/examples/sdk_evolution_agent/collectors.py

@@ -66,7 +66,7 @@ PACKAGE_SOURCE_HINTS: dict[str, tuple[SourceRef, ...]] = {
         SourceRef(
             kind="docs",
             label="OpenAI Codex SDK docs",
-            url="https://developers.openai.com/codex/sdk",
+            url="https://learn.chatgpt.com/docs/codex-sdk",
         ),
     ),
     "openai-codex-cli-bin": (

{agent_runtime_kit-0.5.1 → agent_runtime_kit-0.5.2}/examples/sdk_evolution_agent/models.py

@@ -212,6 +212,9 @@ class RunOptions:
     pr_base: str | None = None
     commit_message: str = "Run SDK evolution update"
     pr_title: str = "Adapt agent-runtime-kit to upstream SDK evolution"
+    model: str | None = None
+    reasoning_effort: str | None = None
+    codex_bin: str | None = None
 
 
 @dataclass(frozen=True)
@@ -226,6 +229,8 @@ class RunContext:
     implementation_enabled: bool
     draft_pr: bool
     event_sink: Any | None = None
+    model: str | None = None
+    reasoning_effort: str | None = None
 
 
 @dataclass(frozen=True)

{agent_runtime_kit-0.5.1 → agent_runtime_kit-0.5.2}/examples/sdk_evolution_agent/release_notes.py

@@ -35,12 +35,12 @@ RELEASE_NOTE_SOURCES: dict[str, tuple[SourceRef, ...]] = {
         SourceRef(
             kind="docs",
             label="Codex SDK docs",
-            url="https://developers.openai.com/codex/sdk",
+            url="https://learn.chatgpt.com/docs/codex-sdk",
         ),
         SourceRef(
             kind="changelog",
             label="Codex changelog",
-            url="https://developers.openai.com/codex/changelog",
+            url="https://learn.chatgpt.com/docs/changelog",
         ),
         SourceRef(
             kind="release",
@@ -349,14 +349,21 @@ def _summaries_for_interval(
 ) -> list[str]:
     text = _prefer_github_discussion_body(text)
     lines = [line.strip() for line in text.splitlines()]
-    version_patterns = [to_version]
+    versions = [to_version]
     if from_version:
-        version_patterns.append(from_version)
+        versions.append(from_version)
+    version_patterns = [
+        re.compile(rf"(?<![\w.+-])v?{re.escape(version)}(?![\w.+-])") for version in versions
+    ]
+    # A baseline mention or a longer version sharing the candidate's prefix is
+    # not release evidence for the candidate itself.
+    if not version_patterns[0].search(text):
+        return []
     matches: list[str] = []
     for index, line in enumerate(lines):
         if not line:
             continue
-        if any(pattern and pattern in line for pattern in version_patterns):
+        if any(pattern.search(line) for pattern in version_patterns):
             matches.append(_clean_summary(line))
             for nearby in lines[index + 1 : index + 9]:
                 cleaned = _clean_summary(nearby)
@@ -364,8 +371,9 @@ def _summaries_for_interval(
                     matches.append(cleaned)
     if not matches and to_version:
         compact = re.sub(r"\s+", " ", text)
-        version_index = compact.find(to_version)
-        if version_index >= 0:
+        version_match = version_patterns[0].search(compact)
+        if version_match:
+            version_index = version_match.start()
             start = max(0, version_index - 160)
             end = min(len(compact), version_index + 320)
             matches.append(_clean_summary(compact[start:end]))
@@ -380,7 +388,12 @@ def _prefer_github_discussion_body(text: str) -> str:
     )
     if not match:
         return text
-    return match.group("body")
+    # GitHub release discussions may identify the version only in the page
+    # title, while the comment body describes the changes without repeating it.
+    # Keep that title attached to its body, excluding navigation and replies.
+    title = re.search(r"<title\b[^>]*>(.*?)</title>", text, flags=re.DOTALL | re.IGNORECASE)
+    body = match.group("body")
+    return f"{title.group(1)}\n{body}" if title else body
 
 
 def _release_note_links_for_version(source_url: str, text: str, to_version: str) -> tuple[str, ...]:

{agent_runtime_kit-0.5.1 → agent_runtime_kit-0.5.2}/examples/sdk_evolution_agent/stages.py

@@ -5,6 +5,7 @@ from __future__ import annotations
 import json
 import re
 from collections.abc import Mapping, Sequence
+from functools import partial
 from pathlib import Path
 from typing import Any
 
@@ -109,7 +110,7 @@ class FixtureEvolutionRuntime:
         await self.aclose()
 
 
-def build_registry() -> RuntimeRegistry:
+def build_registry(*, codex_bin: str | None = None) -> RuntimeRegistry:
     """Build the runtime registry used by the SDK evolution agent."""
 
     registry = create_default_registry(include_fake=False)
@@ -117,7 +118,7 @@ def build_registry() -> RuntimeRegistry:
     register_adapters(registry)
     registry.register(
         AgentRuntimeKind.CODEX_AGENT_SDK,
-        _codex_evolution_runtime,
+        partial(_codex_evolution_runtime, codex_bin=codex_bin),
         replace=True,
     )
     registry.register(
@@ -133,12 +134,16 @@ def build_registry() -> RuntimeRegistry:
     return registry
 
 
-def _codex_evolution_runtime(**kwargs: Any) -> CodexAgentRuntime:
+def _codex_evolution_runtime(*, codex_bin: str | None = None, **kwargs: Any) -> CodexAgentRuntime:
     codex_home = prepare_isolated_codex_home(codex_home=SDK_EVOLUTION_CODEX_HOME)
     env = dict(kwargs.pop("env", {}) or {})
     env.setdefault("CODEX_HOME", str(codex_home))
     kwargs.setdefault("default_model", SDK_EVOLUTION_CODEX_MODEL)
     kwargs.setdefault("reuse_process", True)
+    if codex_bin is not None:
+        from openai_codex import CodexConfig
+
+        kwargs.setdefault("config_cls", partial(CodexConfig, codex_bin=codex_bin))
     return CodexAgentRuntime(env=env, **kwargs)
 
 
@@ -152,10 +157,17 @@ def _antigravity_evolution_runtime(**kwargs: Any) -> AntigravityAgentRuntime:
     return AntigravityAgentRuntime(**kwargs)
 
 
-def resolve_runtime(kind: str, *, registry: RuntimeRegistry | None = None) -> AgentRuntime:
+def resolve_runtime(
+    kind: str, *, registry: RuntimeRegistry | None = None, codex_bin: str | None = None
+) -> AgentRuntime:
     """Resolve a runtime through agent-runtime-kit and verify availability."""
 
-    registry = registry or build_registry()
+    if codex_bin is not None:
+        if kind != "codex-agent-sdk":
+            raise ValueError("codex_bin requires the codex-agent-sdk runtime")
+        if registry is not None:
+            raise ValueError("codex_bin cannot be combined with an injected registry")
+    registry = registry or build_registry(codex_bin=codex_bin)
     runtime = registry.resolve(kind)
     availability = runtime.availability()
     if not availability.available:
@@ -188,6 +200,8 @@ async def run_stage(
         permissions=permissions,
         event_sink=context.event_sink,
         output_schema=schema,
+        model=context.model,
+        reasoning_effort=context.reasoning_effort,
         metadata=_stage_metadata(runtime, stage=stage, context=context),
     )
     try:
@@ -990,8 +1004,10 @@ def _stage_metadata(
 ) -> dict[str, Any]:
     metadata: dict[str, Any] = {"stage": stage, "run_id": context.run_id}
     if runtime.kind is AgentRuntimeKind.CODEX_AGENT_SDK:
-        metadata["model"] = SDK_EVOLUTION_CODEX_MODEL
-        metadata["reasoning_effort"] = SDK_EVOLUTION_CODEX_REASONING_EFFORT
+        metadata["model"] = context.model or SDK_EVOLUTION_CODEX_MODEL
+        metadata["reasoning_effort"] = (
+            context.reasoning_effort or SDK_EVOLUTION_CODEX_REASONING_EFFORT
+        )
     return metadata
 
 

{agent_runtime_kit-0.5.1 → agent_runtime_kit-0.5.2}/pyproject.toml

@@ -4,7 +4,7 @@ build-backend = "hatchling.build"
 
 [project]
 name = "agent-runtime-kit"
-version = "0.5.1"
+version = "0.5.2"
 description = "One typed runtime API for Claude, Codex, and Antigravity agent SDKs."
 readme = "README.md"
 requires-python = ">=3.10"
@@ -48,12 +48,12 @@ dependencies = ["jsonschema>=4.18,<5"]
 claude = ["claude-agent-sdk>=0.2.87,<0.3"]
 # 0.1.0b3 floor: earlier betas pin a codex binary with no manylinux wheels,
 # making the extra uninstallable on glibc Linux.
-codex = ["openai-codex>=0.1.0b3,<0.148"]
+codex = ["openai-codex>=0.1.0b3,<0.155"]
 antigravity = ["google-antigravity>=0.1.2,<0.2"]
 all = [
     "claude-agent-sdk>=0.2.87,<0.3",
     "google-antigravity>=0.1.2,<0.2",
-    "openai-codex>=0.1.0b3,<0.148",
+    "openai-codex>=0.1.0b3,<0.155",
 ]
 
 [project.urls]