alysis-code 0.14.2__tar.gz → 0.14.6__tar.gz

{alysis_code-0.14.2 → alysis_code-0.14.6}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: alysis-code
-Version: 0.14.2
+Version: 0.14.6
 Summary: Local CLI coding agent powered by any API.
 Project-URL: Changelog, https://github.com/AlysisAi/alysis-code/blob/main/docs/CHANGELOG.md
 Project-URL: Documentation, https://github.com/AlysisAi/alysis-code/tree/main/docs
@@ -196,6 +196,8 @@ Persona changes made during a running turn apply after that turn finishes.
 
 Apache-2.0. You're free to use, modify, and distribute this code, including commercially, as long as
 you keep the attribution and license notices. See [LICENSE](LICENSE) and [NOTICE](NOTICE).
+The shared editor sidebar includes MIT-licensed adaptations from Kilo Code/OpenCode; retain its
+[third-party notices](extensions/shared-ui/THIRD_PARTY_NOTICES.txt).
 
 ## FAQ
 
@@ -218,3 +220,11 @@ More questions? See the [full FAQ](https://alysiscode.com/docs/faq).
 ---
 
 **Join the community** — [X](https://x.com/alysiscode) | [GitHub Issues](https://github.com/AlysisAi/alysis-code/issues)
+
+
+## Editor integrations
+
+[VS Code](extensions/vscode-alysis/README.md) uses the shared Alysis sidebar and Python IDE bridge.
+[JetBrains](extensions/jetbrains-alysis/README.md) is a development preview with a smaller feature set.
+The VS Code Marketplace pre-release is pending production signing and acceptance; see the
+[IDE release status](docs/vscode-marketplace-launch.md) before installing component packages.

{alysis_code-0.14.2 → alysis_code-0.14.6}/README.md

@@ -135,6 +135,8 @@ Persona changes made during a running turn apply after that turn finishes.
 
 Apache-2.0. You're free to use, modify, and distribute this code, including commercially, as long as
 you keep the attribution and license notices. See [LICENSE](LICENSE) and [NOTICE](NOTICE).
+The shared editor sidebar includes MIT-licensed adaptations from Kilo Code/OpenCode; retain its
+[third-party notices](extensions/shared-ui/THIRD_PARTY_NOTICES.txt).
 
 ## FAQ
 
@@ -157,3 +159,11 @@ More questions? See the [full FAQ](https://alysiscode.com/docs/faq).
 ---
 
 **Join the community** — [X](https://x.com/alysiscode) | [GitHub Issues](https://github.com/AlysisAi/alysis-code/issues)
+
+
+## Editor integrations
+
+[VS Code](extensions/vscode-alysis/README.md) uses the shared Alysis sidebar and Python IDE bridge.
+[JetBrains](extensions/jetbrains-alysis/README.md) is a development preview with a smaller feature set.
+The VS Code Marketplace pre-release is pending production signing and acceptance; see the
+[IDE release status](docs/vscode-marketplace-launch.md) before installing component packages.

{alysis_code-0.14.2 → alysis_code-0.14.6}/docs/CHANGELOG.md

@@ -1,11 +1,77 @@
 # Changelog
 
+## 0.14.6 — final VSIX attestation and startup validation
+
+- Give each final VSIX dependency report an RFC 4122 document ID accepted by
+  GitHub's attestation parser. Derive it from the exact package bytes so promotion
+  can reproduce the signed report without changing its identity.
+- Wait for startup runtime validation before reporting installed runtime evidence,
+  so a fresh installation does not report the bundled CLI as temporarily unavailable.
+
+All six 0.14.5 native runtimes and their assembled manifest passed signing, but
+final VSIX attestation failed before installation testing. This replacement
+preserves the existing tags; Marketplace pre-release 0.3.0 still requires the
+documented installation, acceptance, and promotion gates.
+
+## 0.14.5 — native managed-runtime packaging
+
+- Pin the six bundled runtimes to checksum-verified CPython 3.12.15 distributions,
+  including their observed OpenSSL and Expat versions in each dependency SBOM.
+- Build Linux executables with shared-library-capable Python on the existing
+  glibc 2.28 baseline, and build tested, static OpenSSL 3.5.9 for Windows ARM64
+  and Intel macOS cryptography.
+- Sign embedded macOS libraries during packaging and verify notarization using
+  Apple's documented check for standalone code. Preserve Windows API paths in Git Bash.
+- Add an unsigned native build and executable smoke check for all six targets.
+
+This source version prepares a replacement candidate after the protected `v0.14.4`
+native build failed. Marketplace pre-release 0.3.0 remains subject to the documented
+signing, installation, acceptance, and promotion gates.
+
+## 0.14.4 — IDE integrations and release tooling
+
+- Add the VS Code 0.3.0 integration with a shared sidebar, provider management, bounded editor
+  context, reviewed changes, recovery checkpoints, and native host actions.
+- Add shared UI transport and a JetBrains development preview using the same Python IDE bridge.
+- Isolate provider credentials, handle physical workspace paths, and keep unit tests portable.
+- Add six-target managed-runtime packaging, Azure OIDC Windows signing verification, Apple
+  notarization/provenance contracts, and protected evidence and Marketplace promotion workflows.
+- Update pinned sandbox build inputs; fresh image scans remain required.
+
+This is a CLI/source release. Production signed VSIX packages and Marketplace publication remain
+pending the [documented signing and acceptance gates](vscode-marketplace-launch.md).
+
 Notable user-facing changes to Alysis Code are recorded here. This project follows
 [Keep a Changelog](https://keepachangelog.com/en/1.1.0/) and uses
 [Semantic Versioning](https://semver.org/spec/v2.0.0.html) while the public API remains pre-1.0.
 
 ## [Unreleased]
 
+## [0.14.3] - 2026-09-29
+
+### Added
+
+- Add Claude Sonnet 5.5 to the Alysis account model picker, using native Messages
+  transport, streaming, signed reasoning and tool continuation, and the shared
+  hosted credit allowance.
+
+### Changed
+
+- Default hosted accounts to automatic prompt caching. Migrate legacy Manual
+  settings only when their other cache settings match the old defaults; preserve
+  Off, customized settings, and subsequent explicit choices.
+- Use five-minute cache controls for hosted Sonnet and provider-managed caching
+  for hosted DeepSeek, GLM, and Luna. See the
+  [caching guide](hosted-prompt-caching.md) for migration details and controls.
+
+### Fixed
+
+- Resolve cache previews and diagnostics using each model's actual protocol,
+  preserve hosted credit-limit error details, and report Sonnet reasoning usage
+  without counting it twice.
+- Honor the configured HTTP MCP startup timeout for initialization acknowledgements.
+  Report MCP startup errors once and exit cleanly without retrying through classic chat.
+
 ## [0.14.2] - 2026-09-28
 
 ### Fixed

{alysis_code-0.14.2 → alysis_code-0.14.6}/pyproject.toml

@@ -4,7 +4,7 @@ build-backend = "hatchling.build"
 
 [project]
 name = "alysis-code"
-version = "0.14.2"
+version = "0.14.6"
 description = "Local CLI coding agent powered by any API."
 readme = "README.md"
 requires-python = ">=3.11"

{alysis_code-0.14.2 → alysis_code-0.14.6}/src/alysis_code/__init__.py

@@ -1,3 +1,3 @@
 __all__ = ["__version__"]
 
-__version__ = "0.14.2"
+__version__ = "0.14.6"

{alysis_code-0.14.2 → alysis_code-0.14.6}/src/alysis_code/cli_impl/chat/loop.py

@@ -18,6 +18,7 @@ from ...cancellation import InteractiveCancellationToken
 from ...compaction.conversation_compactor import CompactionState
 from ...error_text import sanitize_error_text_for_output
 from ...failure_category import exit_code_for_failure
+from ...mcp.errors import McpError
 from ...personas import (
     DEFAULT_PERSONA,
     get_persona,
@@ -3738,6 +3739,10 @@ def chat(
                             )
                         except Exception:
                             pass
+                    if isinstance(_tui_exc, McpError):
+                        # Rebuilding the same MCP session in classic chat cannot
+                        # repair a server startup failure. Report it once below.
+                        raise
                     console.print(
                         f"[yellow]TUI unavailable ({_tui_exc}); using classic chat.[/yellow]"
                     )
@@ -3865,6 +3870,9 @@ def chat(
     except ConfigError as e:
         console.print(f"[red]Config error:[/red] {e}")
         raise typer.Exit(code=2) from e
+    except McpError as e:
+        console.print(f"MCP error: {sanitize_error_text_for_output(e)}", style="red", markup=False)
+        raise typer.Exit(code=1) from e
     except WorkspaceBindingError as e:
         console.print(f"[red]Workspace error:[/red] {e}")
         raise typer.Exit(code=1) from e

{alysis_code-0.14.2 → alysis_code-0.14.6}/src/alysis_code/cli_impl/config_menu.py

@@ -63,6 +63,7 @@ from ..profile_presets import (
 from ..profiles import (
     SUBSCRIPTION_SELECTION_REQUIRED_KEY,
     ProfileSpec,
+    apply_hosted_prompt_cache_defaults,
     get_active_profile,
     list_profiles,
     set_active_profile,
@@ -226,6 +227,10 @@ class ConfigMenuState:
     subscription_selection_required: bool = False
     default_workspace_path: str = ""
     thinking_label_explicitly_set: bool = field(default=False, repr=False)
+    hosted_prompt_cache_defaults_applied: bool = False
+    prompt_cache_mode_explicitly_set: bool = field(default=False, repr=False)
+    cache_prompt_key_enabled: bool = True
+    cache_keepalive_enabled: bool = False
     _subscription_models_cache: tuple[Any, ...] = field(default=(), repr=False)
     _subscription_models_loaded: bool = field(default=False, repr=False)
     _provider_models_cache_identity: tuple[str, ...] | None = field(default=None, repr=False)
@@ -236,6 +241,7 @@ class ConfigMenuState:
 
     @classmethod
     def from_cfg(cls, cfg: AppConfig) -> ConfigMenuState:
+        apply_hosted_prompt_cache_defaults(cfg, get_active_profile(cfg))
         warnings: list[str] = []
         try:
             resolved_key = resolve_api_key(cfg)
@@ -363,6 +369,9 @@ class ConfigMenuState:
             default_workspace_path=default_workspace_path,
         )
         state._normalize_provider_thinking_label()
+        state.hosted_prompt_cache_defaults_applied = cfg.hosted_prompt_cache_defaults_applied
+        state.cache_prompt_key_enabled = cfg.cache.prompt_cache_key_enabled
+        state.cache_keepalive_enabled = cfg.cache.keepalive_enabled
         state._original = state.snapshot()
         return state
 
@@ -394,6 +403,8 @@ class ConfigMenuState:
             "clear_stored_key_confirmed": self.clear_stored_key_confirmed,
             "clear_stored_key_profile": self.clear_stored_key_profile,
             "thinking_label_explicitly_set": self.thinking_label_explicitly_set,
+            "hosted_prompt_cache_defaults_applied": self.hosted_prompt_cache_defaults_applied,
+            "prompt_cache_mode_explicitly_set": self.prompt_cache_mode_explicitly_set,
             "subscription_selection_required": self.subscription_selection_required,
         }
 
@@ -439,6 +450,12 @@ class ConfigMenuState:
         self.thinking_label_explicitly_set = bool(
             self._original.get("thinking_label_explicitly_set", False)
         )
+        self.hosted_prompt_cache_defaults_applied = bool(
+            self._original.get("hosted_prompt_cache_defaults_applied", False)
+        )
+        self.prompt_cache_mode_explicitly_set = bool(
+            self._original.get("prompt_cache_mode_explicitly_set", False)
+        )
         self.refresh_api_key_status()
 
     def set_field(self, name: str, value: str) -> None:
@@ -473,6 +490,16 @@ class ConfigMenuState:
         }:
             raise KeyError(f"Unknown config menu field: {name}")
         self.fields[key] = str(value)
+        if key in {
+            "prompt_cache_mode",
+            "prompt_cache_key",
+            "prompt_cache_retention",
+            "anthropic_prompt_cache_enabled",
+            "anthropic_prompt_cache_ttl",
+        }:
+            self.hosted_prompt_cache_defaults_applied = True
+        if key == "prompt_cache_mode":
+            self.prompt_cache_mode_explicitly_set = True
         if key == "model":
             profile = _active_subscription_profile(self)
             if profile is not None and str(value).strip() != profile.default_model:
@@ -646,6 +673,10 @@ class ConfigMenuState:
         self._provider_models_cache = ()
         self._provider_model_catalog_warning = ""
         profile = ProfileSpec.from_dict(profile_name, self.profiles[profile_name])
+        cache_cfg = self._resolution_cfg()
+        if apply_hosted_prompt_cache_defaults(cache_cfg, profile):
+            self.fields["prompt_cache_mode"] = cache_cfg.prompt_cache_mode
+            self.hosted_prompt_cache_defaults_applied = True
         if profile.base_url:
             self.fields["base_url"] = profile.base_url
         if profile.auth_provider:
@@ -763,6 +794,11 @@ class ConfigMenuState:
             prompt_cache_mode=_normalize_prompt_cache_mode(
                 self.fields.get("prompt_cache_mode", "manual")
             ),
+            hosted_prompt_cache_defaults_applied=self.hosted_prompt_cache_defaults_applied,
+            cache={
+                "prompt_cache_key_enabled": self.cache_prompt_key_enabled,
+                "keepalive_enabled": self.cache_keepalive_enabled,
+            },
             prompt_cache_key=str(self.fields.get("prompt_cache_key", "") or ""),
             prompt_cache_retention=str(self.fields.get("prompt_cache_retention", "") or ""),
             anthropic_prompt_cache_enabled=_normalize_bool_text(
@@ -813,6 +849,7 @@ class ConfigMenuState:
             )
 
         changes: dict[str, Any] = {}
+        cfg.hosted_prompt_cache_defaults_applied = self.hosted_prompt_cache_defaults_applied
 
         execution = getattr(cfg, "execution", None)
         if execution is None:
@@ -930,7 +967,10 @@ class ConfigMenuState:
         desired_cache_mode = _normalize_prompt_cache_mode(
             self.fields.get("prompt_cache_mode", "manual")
         )
-        if str(getattr(cfg, "prompt_cache_mode", "") or "manual") != desired_cache_mode:
+        if (
+            str(getattr(cfg, "prompt_cache_mode", "") or "manual") != desired_cache_mode
+            or self.prompt_cache_mode_explicitly_set
+        ):
             set_config_value(cfg, "prompt_cache_mode", desired_cache_mode)
             changes["prompt_cache_mode"] = desired_cache_mode
 
@@ -1415,8 +1455,12 @@ def _cache_summary_text(state: ConfigMenuState) -> str:
     if mode == "off":
         return f"cache off · {policy_summary} · {compaction}"
     if mode == "auto":
-        ttl = _normalize_anthropic_prompt_cache_ttl(
-            state.fields.get("anthropic_prompt_cache_ttl", "5m")
+        ttl = (
+            policy.anthropic_cache_control_ttl
+            if policy is not None
+            else _normalize_anthropic_prompt_cache_ttl(
+                state.fields.get("anthropic_prompt_cache_ttl", "5m")
+            )
         )
         return f"auto · {policy_summary} · Anthropic TTL {ttl} · {compaction}"
 
@@ -1474,7 +1518,13 @@ def _effective_cache_capability_for_state(
             )
             or preview_profile.name
         )
-        protocol = str(preview_profile.protocol or OPENAI_COMPAT_PROTOCOL).strip()
+        from ..llm.factory import resolve_model_protocol
+
+        protocol = resolve_model_protocol(
+            provider_key=provider_key,
+            model=model,
+            protocol=str(preview_profile.protocol or OPENAI_COMPAT_PROTOCOL).strip(),
+        )
         capabilities = get_provider_protocol_capabilities(
             provider_key=provider_key,
             protocol=protocol,
@@ -1532,7 +1582,13 @@ def _resolved_cache_policy_with_error_for_state(
             )
             or profile.name
         )
-        protocol = str(profile.protocol or OPENAI_COMPAT_PROTOCOL).strip()
+        from ..llm.factory import resolve_model_protocol
+
+        protocol = resolve_model_protocol(
+            provider_key=provider_key,
+            model=model,
+            protocol=str(profile.protocol or OPENAI_COMPAT_PROTOCOL).strip(),
+        )
         capabilities = get_provider_protocol_capabilities(
             provider_key=provider_key,
             protocol=protocol,

{alysis_code-0.14.2 → alysis_code-0.14.6}/src/alysis_code/cli_impl/tui/config_flow.py

@@ -406,9 +406,9 @@ _THINKING_DESCRIPTIONS = {
     "auto": "let the provider decide",
 }
 _CACHE_MODE_DESCRIPTIONS = {
-    "auto": "derive provider-aware cache settings when supported",
+    "auto": "use supported provider caching; default for Alysis hosted models",
     "manual": "use the manual key/retention below when supported",
-    "off": "disable prompt caching",
+    "off": "disable optional client cache controls",
 }
 _CACHE_TTL_DESCRIPTIONS = {
     "5m": "lower lifetime; safest default",

{alysis_code-0.14.2 → alysis_code-0.14.6}/src/alysis_code/config.py

@@ -699,6 +699,9 @@ class AppConfig(BaseModel):
     session_log_dir: str | None = None
     crash_diagnostic_log_path: str | None = None
     prompt_cache_mode: str = "manual"
+    # Hosted defaults are applied once when an Alysis profile is selected/loaded.
+    # Explicit cache choices also set this marker, including reselecting Manual.
+    hosted_prompt_cache_defaults_applied: bool = False
     prompt_cache_key: str | None = None
     prompt_cache_retention: str | None = None
     anthropic_prompt_cache_enabled: bool = False
@@ -1508,6 +1511,9 @@ def resolve_profile_api_key(
     profile = get_profile(cfg, profile_name)
     if profile is None:
         return ApiKeyResolution(key=None, source="missing")
+    ide_key = _resolve_ide_profile_api_key(cfg, profile_name)
+    if ide_key.key:
+        return ide_key
     stored_profile_key = _resolve_stored_profile_api_key(cfg, profile_name)
     if stored_profile_key.key:
         return stored_profile_key
@@ -1528,6 +1534,21 @@ def resolve_profile_api_key(
     return ApiKeyResolution(key=None, source="missing")
 
 
+def _resolve_ide_profile_api_key(cfg: AppConfig, profile_name: str) -> ApiKeyResolution:
+    """Honor an IDE's explicitly supplied profile key without overwriting CLI credentials."""
+    from .profiles import get_profile
+
+    profile = get_profile(cfg, profile_name)
+    if profile is not None:
+        # Match the extension's encoding; plain env-var names can collide between two profiles.
+        encoded = profile.name.encode("utf-8").hex().upper()
+        env_name = f"ALYSIS_IDE_PROFILE_{encoded}_API_KEY"
+        key = str(os.environ.get(env_name) or "").strip()
+        if key:
+            return ApiKeyResolution(key=key, source=f"env:{env_name}")
+    return ApiKeyResolution(key=None, source="missing")
+
+
 def _resolve_stored_profile_api_key(cfg: AppConfig, profile_name: str) -> ApiKeyResolution:
     from .profiles import get_profile
 
@@ -1549,6 +1570,9 @@ def resolve_api_key(
     if profile_name is None:
         profile_name = str((effective_cfg.extra_fields or {}).get("active_profile") or "").strip()
     if profile_name:
+        ide_key = _resolve_ide_profile_api_key(effective_cfg, profile_name)
+        if ide_key.key:
+            return ide_key
         stored_profile_key = _resolve_stored_profile_api_key(effective_cfg, profile_name)
         if stored_profile_key.key:
             return stored_profile_key
@@ -3624,22 +3648,27 @@ def set_config_value(
 
     if key == "prompt_cache_mode":
         cfg.prompt_cache_mode = _normalize_prompt_cache_mode(value, key=key)
+        cfg.hosted_prompt_cache_defaults_applied = True
         return cfg
 
     if key == "prompt_cache_key":
         cfg.prompt_cache_key = value.strip() or None
+        cfg.hosted_prompt_cache_defaults_applied = True
         return cfg
 
     if key == "prompt_cache_retention":
         cfg.prompt_cache_retention = value.strip() or None
+        cfg.hosted_prompt_cache_defaults_applied = True
         return cfg
 
     if key == "anthropic_prompt_cache_enabled":
         cfg.anthropic_prompt_cache_enabled = _coerce_bool(value, key=key)
+        cfg.hosted_prompt_cache_defaults_applied = True
         return cfg
 
     if key == "anthropic_prompt_cache_ttl":
         cfg.anthropic_prompt_cache_ttl = _normalize_anthropic_prompt_cache_ttl(value, key=key)
+        cfg.hosted_prompt_cache_defaults_applied = True
         return cfg
 
     if key == "verify_commands":

{alysis_code-0.14.2 → alysis_code-0.14.6}/src/alysis_code/git_evidence.py

@@ -11,7 +11,7 @@ from typing import Any, Literal
 
 from .file_classification import classify_path, is_generated_or_vendor_path
 from .git_ops import GitOpsError
-from .git_safe import build_git_cmd, build_git_process_env
+from .git_safe import build_git_cmd, build_git_process_env, configured_git_filter_drivers
 from .runtime_artifacts import is_runtime_artifact_path
 
 PatchCaptureStatus = Literal["complete", "partial", "failed"]
@@ -445,11 +445,13 @@ def _snapshot_git(
     args: list[str],
     *,
     env: dict[str, str] | None = None,
+    input_text: str | None = None,
 ) -> str:
     try:
         result = subprocess.run(
             build_git_cmd(root, args, env=env, disable_filters=True),
             env=build_git_process_env(env),
+            input=input_text.encode("utf-8") if input_text is not None else None,
             capture_output=True,
             check=False,
         )
@@ -504,17 +506,63 @@ def _check_snapshot_files(root: Path, *, excluded_paths: tuple[str, ...]) -> Non
         raise WorkspaceSnapshotError("unsupported_workspace_file", paths=tuple(sorted(unsupported)))
 
 
+def _check_snapshot_filters(root: Path, *, pathspecs: list[str]) -> None:
+    # Git status may recognize a size change without converting file content.
+    # Resolve attributes against the actual index before a HEAD-based temporary
+    # index can lose an index-only .gitattributes fallback.
+    try:
+        drivers = configured_git_filter_drivers(root)
+    except (OSError, subprocess.TimeoutExpired) as exc:
+        raise WorkspaceSnapshotError("workspace_snapshot_unreadable") from exc
+    if not drivers:
+        return
+    paths = _snapshot_git(
+        root,
+        [
+            "--no-optional-locks",
+            "ls-files",
+            "--cached",
+            "--others",
+            "--exclude-standard",
+            "-z",
+            "--",
+            *pathspecs,
+        ],
+    )
+    attributes = _snapshot_git(
+        root, ["check-attr", "-z", "--stdin", "filter"], input_text=paths
+    ).split("\0")
+    if attributes[-1] != "" or (len(attributes) - 1) % 3:
+        raise WorkspaceSnapshotError("workspace_snapshot_unreadable")
+    affected = tuple(
+        sorted(
+            {
+                attributes[index]
+                for index in range(0, len(attributes) - 1, 3)
+                if attributes[index + 2] in drivers
+            }
+        )
+    )
+    if affected:
+        raise WorkspaceSnapshotError(
+            "workspace_snapshot_git_failed",
+            paths=affected,
+            detail="Git snapshot refused: external Git filters are disabled for isolated workspaces",
+        )
+
+
 def _working_tree_snapshot(
     root: Path, *, base_ref: str, excluded_paths: tuple[str, ...] = ()
 ) -> str:
     _check_snapshot_files(root, excluded_paths=excluded_paths)
+    pathspecs = [".", *(f":(top,exclude,literal){path}" for path in excluded_paths)]
+    _check_snapshot_filters(root, pathspecs=pathspecs)
     # A separate index preserves both staged and unstaged parent work. Git handles
     # binary data, modes, names and ignored paths; no filename classifier is used.
     with tempfile.TemporaryDirectory(prefix="alysis-git-snapshot-") as scratch:
         env = build_git_process_env()
         env["GIT_INDEX_FILE"] = os.fspath(Path(scratch) / "index")
         _snapshot_git(root, ["read-tree", base_ref], env=env)
-        pathspecs = [".", *(f":(top,exclude,literal){path}" for path in excluded_paths)]
         _snapshot_git(root, ["add", "--all", "--", *pathspecs], env=env)
         tree = _snapshot_git(root, ["write-tree"], env=env).strip()
         gitlinks = tuple(

{alysis_code-0.14.2 → alysis_code-0.14.6}/src/alysis_code/git_safe.py

@@ -77,11 +77,29 @@ def _disabled_filter_config(
     returned command, so unused drivers do not prevent ordinary repositories from
     working. Re-read for every command: a child can introduce filter configuration.
     """
+    drivers = configured_git_filter_drivers(root, extra_config=config, env=env)
+    overrides: dict[str, str] = {}
+    for driver in drivers:
+        # Empty all three commands, including the long-running process protocol.
+        # required=true makes an attempted conversion fail instead of passing
+        # through bytes that might differ from the real filter's representation.
+        overrides.update({f"filter.{driver}.{kind}": "" for kind in ("clean", "smudge", "process")})
+        overrides[f"filter.{driver}.required"] = "true"
+    return overrides
+
+
+def configured_git_filter_drivers(
+    root: Path,
+    *,
+    extra_config: dict[str, str] | None = None,
+    env: Mapping[str, str] | None = None,
+) -> tuple[str, ...]:
+    """Read effective nonempty filter commands without invoking any driver."""
     result = subprocess.run(
         build_git_cmd(
             root,
             ["config", "--null", "--get-regexp", r"^filter\..*\.(clean|smudge|process)$"],
-            extra_config=config,
+            extra_config=extra_config,
             env=env,
         ),
         env=build_git_process_env(env),
@@ -90,7 +108,7 @@ def _disabled_filter_config(
         timeout=5,
     )
     if result.returncode == 1:  # No configured filter commands.
-        return {}
+        return ()
     if result.returncode != 0:
         raise OSError("Could not inspect Git filter configuration")
     values: dict[str, str] = {}
@@ -100,12 +118,6 @@ def _disabled_filter_config(
             if not separator:
                 raise OSError("Could not read Git filter configuration")
             values[key] = value
-    drivers = {key.rsplit(".", 1)[0] for key, value in values.items() if value}
-    overrides: dict[str, str] = {}
-    for driver in sorted(drivers):
-        # Empty all three commands, including the long-running process protocol.
-        # required=true makes an attempted conversion fail instead of passing
-        # through bytes that might differ from the real filter's representation.
-        overrides.update({f"{driver}.{kind}": "" for kind in ("clean", "smudge", "process")})
-        overrides[f"{driver}.required"] = "true"
-    return overrides
+    return tuple(
+        sorted({key[len("filter.") :].rsplit(".", 1)[0] for key, value in values.items() if value})
+    )

{alysis_code-0.14.2 → alysis_code-0.14.6}/src/alysis_code/ide/activity_events.py

@@ -25,7 +25,7 @@ ActivityKind: TypeAlias = Literal[
     "other",
 ]
 ActivityStatus: TypeAlias = Literal[
-    "queued", "running", "succeeded", "failed", "cancelled", "blocked"
+    "queued", "running", "succeeded", "failed", "cancelled", "blocked", "completed_unverified"
 ]
 
 MAX_ACTIVITY_TEXT_CHARS = 500
@@ -47,7 +47,15 @@ ACTIVITY_KINDS = (
     "plan",
     "other",
 )
-ACTIVITY_STATUSES = ("queued", "running", "succeeded", "failed", "cancelled", "blocked")
+ACTIVITY_STATUSES = (
+    "queued",
+    "running",
+    "succeeded",
+    "failed",
+    "cancelled",
+    "blocked",
+    "completed_unverified",
+)
 
 _SAFE_METADATA_KEYS = frozenset(
     {
@@ -64,6 +72,8 @@ _SAFE_METADATA_KEYS = frozenset(
         "step",
         "truncated",
         "worker_id",
+        "task_outcome",
+        "verified_success",
     }
 )
 _INTERNAL_NAME_PATTERN = re.compile(r"(?:^|[._:/-])(?:rs|internal)(?:[._:/-]|$)", re.I)
@@ -335,6 +345,8 @@ def _normalize_status(status: str) -> ActivityStatus:
         return "queued"
     if normalized in {"running", "started", "in_progress", "progress"}:
         return "running"
+    if normalized == "completed_unverified":
+        return "completed_unverified"
     if normalized in {"done", "completed", "complete", "success", "succeeded", "ok"}:
         return "succeeded"
     if normalized in {"cancelled", "canceled", "interrupted"}:

{alysis_code-0.14.2 → alysis_code-0.14.6}/src/alysis_code/ide/artifacts.py

@@ -4,6 +4,7 @@ from dataclasses import dataclass
 from pathlib import Path, PurePosixPath
 from typing import Any
 
+from ..session_artifacts import ARTIFACT_HANDLE_INDEX_DIR
 from .protocol import ProtocolError, redact_secrets
 
 DEFAULT_MAX_ARTIFACT_BYTES = 64 * 1024
@@ -50,6 +51,8 @@ class ArtifactStore:
                 except OSError:
                     continue
                 for path in entries:
+                    if depth == 0 and path.name == ARTIFACT_HANDLE_INDEX_DIR:
+                        continue
                     if path.is_symlink():
                         continue
                     if path.is_dir():