cavi-mcp-core 0.1.0__tar.gz

This diff represents the content of publicly available package versions that have been released to one of the supported registries. The information contained in this diff is provided for informational purposes only and reflects changes between package versions as they appear in their respective public registries.
@@ -0,0 +1,10 @@
1
+ .venv/
2
+ __pycache__/
3
+ *.pyc
4
+ .ruff_cache/
5
+ dist/
6
+ build/
7
+ *.egg-info/
8
+ .tmp/
9
+ .worktrees/
10
+ .DS_Store
@@ -0,0 +1,13 @@
1
+ # Changelog
2
+
3
+ ## Unreleased
4
+
5
+ ## 0.1.0 - 2026-11-04
6
+
7
+ ### Added
8
+
9
+ - `Server` with compact tool schemas, anticipated-error reporting and one structured failure for every tool error, including unknown tools.
10
+ - `Result`, `VariantResult`, `OperationRejected`, `OperationUnknown` and `failure` result contracts.
11
+ - Schema compaction and variant validation helpers.
12
+ - Bounded subprocess runner with process-group termination.
13
+ - Workspace-confined bounded file reads and an atomic JSON operation journal.
@@ -0,0 +1,21 @@
1
+ MIT License
2
+
3
+ Copyright (c) 2026 Sasan Sotoodehfar
4
+
5
+ Permission is hereby granted, free of charge, to any person obtaining a copy
6
+ of this software and associated documentation files (the "Software"), to deal
7
+ in the Software without restriction, including without limitation the rights
8
+ to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
9
+ copies of the Software, and to permit persons to whom the Software is
10
+ furnished to do so, subject to the following conditions:
11
+
12
+ The above copyright notice and this permission notice shall be included in all
13
+ copies or substantial portions of the Software.
14
+
15
+ THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
16
+ IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
17
+ FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
18
+ AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
19
+ LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
20
+ OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
21
+ SOFTWARE.
@@ -0,0 +1,123 @@
1
+ Metadata-Version: 2.5
2
+ Name: cavi-mcp-core
3
+ Version: 0.1.0
4
+ Summary: Server type, result contracts, schemas, bounded processes and file journals for CAVI MCP servers
5
+ Project-URL: Homepage, https://github.com/cavi-ai/mcp-core#readme
6
+ Project-URL: Repository, https://github.com/cavi-ai/mcp-core
7
+ Project-URL: Issues, https://github.com/cavi-ai/mcp-core/issues
8
+ Project-URL: Changelog, https://github.com/cavi-ai/mcp-core/blob/main/CHANGELOG.md
9
+ Author: CAVI AI
10
+ License-Expression: MIT
11
+ License-File: LICENSE
12
+ Keywords: mcp,mcp-server,model-context-protocol
13
+ Classifier: Development Status :: 3 - Alpha
14
+ Classifier: Intended Audience :: Developers
15
+ Classifier: Operating System :: MacOS
16
+ Classifier: Operating System :: Microsoft :: Windows
17
+ Classifier: Operating System :: POSIX :: Linux
18
+ Classifier: Programming Language :: Python :: 3
19
+ Classifier: Programming Language :: Python :: 3.11
20
+ Classifier: Programming Language :: Python :: 3.12
21
+ Classifier: Programming Language :: Python :: 3.13
22
+ Classifier: Topic :: Software Development :: Libraries
23
+ Requires-Python: >=3.11
24
+ Requires-Dist: mcp<3,>=2.3
25
+ Requires-Dist: pydantic<3,>=2.11
26
+ Description-Content-Type: text/markdown
27
+
28
+ # Cavi MCP Core
29
+
30
+ Shared plumbing for the CAVI MCP servers ([blender-mcp](https://github.com/cavi-ai/blender-mcp),
31
+ [substance-mcp](https://github.com/cavi-ai/substance-mcp),
32
+ [houdini-mcp](https://github.com/cavi-ai/houdini-mcp),
33
+ [adobe-mcp](https://github.com/cavi-ai/adobe-mcp)): one server type, one failure contract,
34
+ one schema compaction, one bounded process runner, bounded file reads and an operation
35
+ journal.
36
+
37
+ ## Install
38
+
39
+ ```sh
40
+ uv add cavi-mcp-core
41
+ ```
42
+
43
+ Requires Python 3.11+ and the MCP Python SDK 2.
44
+
45
+ ## Server
46
+
47
+ ```python
48
+ from cavi_mcp_core.server import READ_ONLY, Server
49
+
50
+ mcp = Server("my-mcp", package="my-mcp", instructions="Inspect before editing.")
51
+
52
+
53
+ @mcp.tool(annotations=READ_ONLY)
54
+ async def status() -> dict:
55
+ return {"ready": True}
56
+ ```
57
+
58
+ `Server` extends `mcp.server.mcpserver.MCPServer`:
59
+
60
+ - `package` sets the reported server version from that distribution's metadata.
61
+ - Tool schemas are compacted (`compact=False` keeps the SDK's schemas).
62
+ - `ValueError`, `RuntimeError` and `OSError` raised by a tool reach the caller with their
63
+ text; any other exception is logged on the server and reported without details. Pass
64
+ `anticipated=()` to leave tool errors to the SDK.
65
+ - Every tool error, including an unknown tool name, returns one structured result:
66
+
67
+ ```json
68
+ {"error": {"version": 1, "code": "INVALID_ARGUMENT", "operation": "status",
69
+ "outcome": "rejected", "retry_safe": false, "next_action": "inspect_inputs"}}
70
+ ```
71
+
72
+ `outcome` is `rejected` only for invalid arguments, missing inputs and unknown tools;
73
+ anything else is `unknown`, and `retry_safe` is always false. Raise
74
+ `contracts.OperationRejected(message, code)` or `contracts.OperationUnknown(message, code)`
75
+ to report a specific code. Override `Server.failure` to name a different recovery tool
76
+ for unknown outcomes.
77
+
78
+ `need(value, name, action)` and `reject_foreign(action, given, allowed)` validate
79
+ action-style arguments. `READ_ONLY`, `WRITES` and `DESTRUCTIVE` are the tool annotations.
80
+
81
+ ## Contracts
82
+
83
+ `contracts.Result` is a strict pydantic model whose unset optional fields stay off the
84
+ wire. `contracts.VariantResult` accepts one of several declared field sets (`shapes`)
85
+ without a discriminator field.
86
+
87
+ ## Schemas
88
+
89
+ `schema.compact_schema(schema, output=False)` returns a copy without titles or
90
+ `additionalProperties: true`, with equivalent union forms shortened, and validates the same
91
+ values; `output=True` also drops defaults, which are documentation only in a response.
92
+ `schema.validate_variant` and `schema.variant_schema` back `VariantResult`.
93
+
94
+ ## Processes
95
+
96
+ `proc.run(args, timeout)` runs a command with bounded output and one deadline. On a
97
+ timeout or a cancellation it stops the whole process group (process tree on Windows) and
98
+ returns once the process exits. `proc.run_sync` is the blocking form and
99
+ `proc.diagnostics(result)` formats what a failed run printed.
100
+
101
+ ## Files and journals
102
+
103
+ `files.inside(workspace, path)` resolves a path inside an existing workspace and refuses
104
+ `..`, escapes through symbolic links and the workspace root. `files.read_bytes`,
105
+ `files.load_json` and `files.digest` read bounded regular files and refuse a file that
106
+ changes while it is read; `files.parse_json` refuses NaN and Infinity.
107
+
108
+ `journal.write_json` replaces a JSON file atomically, durably and readable only by its
109
+ owner. `journal.completion` reads the completion receipt of one operation,
110
+ `journal.archive` records how an interrupted operation was resolved, and
111
+ `journal.fingerprint` hashes canonical JSON.
112
+
113
+ ## Development
114
+
115
+ ```sh
116
+ uv sync --locked --dev
117
+ uv run --locked python -m unittest discover -s tests
118
+ uv run --locked ruff check .
119
+ ```
120
+
121
+ ## License
122
+
123
+ MIT
@@ -0,0 +1,96 @@
1
+ # Cavi MCP Core
2
+
3
+ Shared plumbing for the CAVI MCP servers ([blender-mcp](https://github.com/cavi-ai/blender-mcp),
4
+ [substance-mcp](https://github.com/cavi-ai/substance-mcp),
5
+ [houdini-mcp](https://github.com/cavi-ai/houdini-mcp),
6
+ [adobe-mcp](https://github.com/cavi-ai/adobe-mcp)): one server type, one failure contract,
7
+ one schema compaction, one bounded process runner, bounded file reads and an operation
8
+ journal.
9
+
10
+ ## Install
11
+
12
+ ```sh
13
+ uv add cavi-mcp-core
14
+ ```
15
+
16
+ Requires Python 3.11+ and the MCP Python SDK 2.
17
+
18
+ ## Server
19
+
20
+ ```python
21
+ from cavi_mcp_core.server import READ_ONLY, Server
22
+
23
+ mcp = Server("my-mcp", package="my-mcp", instructions="Inspect before editing.")
24
+
25
+
26
+ @mcp.tool(annotations=READ_ONLY)
27
+ async def status() -> dict:
28
+ return {"ready": True}
29
+ ```
30
+
31
+ `Server` extends `mcp.server.mcpserver.MCPServer`:
32
+
33
+ - `package` sets the reported server version from that distribution's metadata.
34
+ - Tool schemas are compacted (`compact=False` keeps the SDK's schemas).
35
+ - `ValueError`, `RuntimeError` and `OSError` raised by a tool reach the caller with their
36
+ text; any other exception is logged on the server and reported without details. Pass
37
+ `anticipated=()` to leave tool errors to the SDK.
38
+ - Every tool error, including an unknown tool name, returns one structured result:
39
+
40
+ ```json
41
+ {"error": {"version": 1, "code": "INVALID_ARGUMENT", "operation": "status",
42
+ "outcome": "rejected", "retry_safe": false, "next_action": "inspect_inputs"}}
43
+ ```
44
+
45
+ `outcome` is `rejected` only for invalid arguments, missing inputs and unknown tools;
46
+ anything else is `unknown`, and `retry_safe` is always false. Raise
47
+ `contracts.OperationRejected(message, code)` or `contracts.OperationUnknown(message, code)`
48
+ to report a specific code. Override `Server.failure` to name a different recovery tool
49
+ for unknown outcomes.
50
+
51
+ `need(value, name, action)` and `reject_foreign(action, given, allowed)` validate
52
+ action-style arguments. `READ_ONLY`, `WRITES` and `DESTRUCTIVE` are the tool annotations.
53
+
54
+ ## Contracts
55
+
56
+ `contracts.Result` is a strict pydantic model whose unset optional fields stay off the
57
+ wire. `contracts.VariantResult` accepts one of several declared field sets (`shapes`)
58
+ without a discriminator field.
59
+
60
+ ## Schemas
61
+
62
+ `schema.compact_schema(schema, output=False)` returns a copy without titles or
63
+ `additionalProperties: true`, with equivalent union forms shortened, and validates the same
64
+ values; `output=True` also drops defaults, which are documentation only in a response.
65
+ `schema.validate_variant` and `schema.variant_schema` back `VariantResult`.
66
+
67
+ ## Processes
68
+
69
+ `proc.run(args, timeout)` runs a command with bounded output and one deadline. On a
70
+ timeout or a cancellation it stops the whole process group (process tree on Windows) and
71
+ returns once the process exits. `proc.run_sync` is the blocking form and
72
+ `proc.diagnostics(result)` formats what a failed run printed.
73
+
74
+ ## Files and journals
75
+
76
+ `files.inside(workspace, path)` resolves a path inside an existing workspace and refuses
77
+ `..`, escapes through symbolic links and the workspace root. `files.read_bytes`,
78
+ `files.load_json` and `files.digest` read bounded regular files and refuse a file that
79
+ changes while it is read; `files.parse_json` refuses NaN and Infinity.
80
+
81
+ `journal.write_json` replaces a JSON file atomically, durably and readable only by its
82
+ owner. `journal.completion` reads the completion receipt of one operation,
83
+ `journal.archive` records how an interrupted operation was resolved, and
84
+ `journal.fingerprint` hashes canonical JSON.
85
+
86
+ ## Development
87
+
88
+ ```sh
89
+ uv sync --locked --dev
90
+ uv run --locked python -m unittest discover -s tests
91
+ uv run --locked ruff check .
92
+ ```
93
+
94
+ ## License
95
+
96
+ MIT
@@ -0,0 +1,49 @@
1
+ [build-system]
2
+ requires = ["hatchling>=1.27"]
3
+ build-backend = "hatchling.build"
4
+
5
+ [project]
6
+ name = "cavi-mcp-core"
7
+ version = "0.1.0"
8
+ description = "Server type, result contracts, schemas, bounded processes and file journals for CAVI MCP servers"
9
+ readme = "README.md"
10
+ license = "MIT"
11
+ license-files = ["LICENSE"]
12
+ authors = [{ name = "CAVI AI" }]
13
+ keywords = ["mcp", "model-context-protocol", "mcp-server"]
14
+ classifiers = [
15
+ "Development Status :: 3 - Alpha",
16
+ "Intended Audience :: Developers",
17
+ "Operating System :: MacOS",
18
+ "Operating System :: Microsoft :: Windows",
19
+ "Operating System :: POSIX :: Linux",
20
+ "Programming Language :: Python :: 3",
21
+ "Programming Language :: Python :: 3.11",
22
+ "Programming Language :: Python :: 3.12",
23
+ "Programming Language :: Python :: 3.13",
24
+ "Topic :: Software Development :: Libraries",
25
+ ]
26
+ requires-python = ">=3.11"
27
+ dependencies = ["mcp>=2.3,<3", "pydantic>=2.11,<3"]
28
+
29
+ [project.urls]
30
+ Homepage = "https://github.com/cavi-ai/mcp-core#readme"
31
+ Repository = "https://github.com/cavi-ai/mcp-core"
32
+ Issues = "https://github.com/cavi-ai/mcp-core/issues"
33
+ Changelog = "https://github.com/cavi-ai/mcp-core/blob/main/CHANGELOG.md"
34
+
35
+ [dependency-groups]
36
+ dev = ["ruff>=0.14,<1"]
37
+
38
+ [tool.hatch.build.targets.wheel]
39
+ packages = ["src/cavi_mcp_core"]
40
+
41
+ [tool.hatch.build.targets.sdist]
42
+ include = ["/src", "/tests", "/README.md", "/LICENSE", "/CHANGELOG.md", "/pyproject.toml"]
43
+
44
+ [tool.ruff]
45
+ line-length = 100
46
+ target-version = "py311"
47
+
48
+ [tool.ruff.lint]
49
+ select = ["E", "F", "I", "B", "UP", "SIM"]
@@ -0,0 +1 @@
1
+ """Shared plumbing for CAVI MCP servers: server type, result contracts, schemas, processes."""
@@ -0,0 +1,105 @@
1
+ """Response models and the structured failure every tool error reports."""
2
+
3
+ from typing import ClassVar
4
+
5
+ from mcp.server.mcpserver.exceptions import ToolError, UnexpectedToolError
6
+ from mcp.types import CallToolResult, TextContent
7
+ from pydantic import BaseModel, ConfigDict, ValidationError, model_serializer, model_validator
8
+
9
+ from .schema import validate_variant, variant_schema
10
+
11
+
12
+ class Result(BaseModel):
13
+ """A tool result. Optional fields the tool did not set stay off the wire."""
14
+
15
+ model_config = ConfigDict(extra="allow", strict=True)
16
+
17
+ @model_serializer(mode="wrap")
18
+ def preserve_shape(self, handler):
19
+ present = set(self.model_fields_set)
20
+ for name in self.model_fields_set:
21
+ field = type(self).model_fields.get(name)
22
+ if field is not None and field.alias:
23
+ present.add(field.alias)
24
+ return {key: value for key, value in handler(self).items() if key in present}
25
+
26
+
27
+ class VariantResult(Result):
28
+ """A result with several valid shapes, validated without adding a discriminator field."""
29
+
30
+ shapes: ClassVar[tuple[tuple[str, ...], ...]] = ()
31
+ nullable_shape_fields: ClassVar[frozenset[str]] = frozenset()
32
+
33
+ @model_validator(mode="before")
34
+ @classmethod
35
+ def require_shape(cls, value):
36
+ fields = {field.alias or name for name, field in cls.model_fields.items()}
37
+ return validate_variant(value, fields, cls.shapes, cls.nullable_shape_fields)
38
+
39
+ @classmethod
40
+ def __get_pydantic_json_schema__(cls, core_schema, handler):
41
+ return variant_schema(core_schema, handler, cls.shapes, cls.nullable_shape_fields)
42
+
43
+
44
+ class UnknownToolError(ToolError):
45
+ """The requested name was absent before any tool execution."""
46
+
47
+
48
+ class OperationRejected(ValueError):
49
+ """Refused before any effect; ``error_code`` names the reason."""
50
+
51
+ def __init__(self, message, code="INVALID_ARGUMENT"):
52
+ super().__init__(message)
53
+ self.error_code = code
54
+
55
+
56
+ class OperationUnknown(RuntimeError):
57
+ """The effect may or may not have happened; inspect before any retry."""
58
+
59
+ def __init__(self, message, code="OUTCOME_UNKNOWN"):
60
+ super().__init__(message)
61
+ self.error_code = code
62
+
63
+
64
+ def failure(name, error, unknown_next_action="inspect_application_and_outputs"):
65
+ """The error result for ``error`` raised by tool ``name``; never authorizes a retry.
66
+
67
+ Only local validation, a missing input or a missing tool counts as rejected. Anything
68
+ else leaves the outcome unknown and points the caller at ``unknown_next_action``.
69
+ """
70
+ cause = error
71
+ while (isinstance(cause, ToolError) and not isinstance(cause, UnexpectedToolError)
72
+ and cause.__cause__ is not None):
73
+ cause = cause.__cause__
74
+ missing_tool = isinstance(cause, UnknownToolError)
75
+ rejected = missing_tool or isinstance(cause, (ValueError, ValidationError, FileNotFoundError))
76
+ if isinstance(cause, (OperationRejected, OperationUnknown)):
77
+ code = cause.error_code
78
+ else:
79
+ code = (
80
+ "TOOL_NOT_FOUND"
81
+ if missing_tool
82
+ else "INVALID_ARGUMENT"
83
+ if isinstance(cause, (ValueError, ValidationError))
84
+ else "INPUT_NOT_FOUND"
85
+ if isinstance(cause, FileNotFoundError)
86
+ else "OUTCOME_UNKNOWN"
87
+ )
88
+ return CallToolResult(
89
+ isError=True,
90
+ content=[TextContent(type="text", text=str(error))],
91
+ structuredContent={
92
+ "error": {
93
+ "version": 1,
94
+ "code": code,
95
+ "operation": name,
96
+ "outcome": "rejected" if rejected else "unknown",
97
+ "retry_safe": False,
98
+ "next_action": "list_tools"
99
+ if missing_tool
100
+ else "inspect_inputs"
101
+ if rejected
102
+ else unknown_next_action,
103
+ }
104
+ },
105
+ )
@@ -0,0 +1,56 @@
1
+ """Bounded, workspace-confined file reads shared by local tool workflows."""
2
+
3
+ import hashlib
4
+ import json
5
+ import os
6
+ import stat
7
+ from pathlib import Path
8
+
9
+
10
+ def inside(workspace, value):
11
+ """``value`` resolved inside ``workspace``; refuses traversal, escapes and the root itself."""
12
+ root = Path(workspace).resolve(strict=True)
13
+ path = Path(value)
14
+ if '..' in path.parts:
15
+ raise ValueError('Use workspace paths without parent traversal')
16
+ logical = Path(os.path.abspath(root / path))
17
+ resolved = logical.resolve()
18
+ if (not logical.is_relative_to(root) or not resolved.is_relative_to(root)
19
+ or resolved == root):
20
+ raise ValueError('Path must remain inside the shared workspace')
21
+ return resolved
22
+
23
+
24
+ def read_bytes(path, limit):
25
+ """A regular file's bytes, at most ``limit``; refuses a file that changes while read."""
26
+ descriptor = os.open(path, os.O_RDONLY | getattr(os, 'O_NONBLOCK', 0))
27
+ with os.fdopen(descriptor, 'rb') as stream:
28
+ before = os.fstat(stream.fileno())
29
+ if not stat.S_ISREG(before.st_mode) or before.st_size > limit:
30
+ raise ValueError('Expected a bounded regular file')
31
+ data = stream.read(limit + 1)
32
+ after = os.fstat(stream.fileno())
33
+ current = Path(path).stat()
34
+ if (len(data) > limit or (before.st_size, before.st_mtime_ns, before.st_ino)
35
+ != (after.st_size, after.st_mtime_ns, after.st_ino)
36
+ or (current.st_size, current.st_mtime_ns, current.st_ino)
37
+ != (after.st_size, after.st_mtime_ns, after.st_ino)):
38
+ raise ValueError('Input changed while reading')
39
+ return data
40
+
41
+
42
+ def digest(path):
43
+ """SHA-256 of a regular file of at most 64 MiB."""
44
+ return hashlib.sha256(read_bytes(path, 64 * 1024 * 1024)).hexdigest()
45
+
46
+
47
+ def parse_json(raw):
48
+ """JSON without NaN or Infinity."""
49
+ def invalid(value):
50
+ raise ValueError('Non-finite JSON number: ' + value)
51
+ return json.loads(raw, parse_constant=invalid)
52
+
53
+
54
+ def load_json(path):
55
+ """``parse_json`` of a regular file of at most 1 MiB."""
56
+ return parse_json(read_bytes(path, 1024 * 1024))
@@ -0,0 +1,66 @@
1
+ """Durable JSON state and retained native completion receipts."""
2
+
3
+ import hashlib
4
+ import json
5
+ import os
6
+ import uuid
7
+ from pathlib import Path
8
+
9
+
10
+ def write_json(path, value):
11
+ """Replace ``path`` with ``value`` atomically and durably; the file is private to the user."""
12
+ path = Path(path)
13
+ temporary = path.with_name("." + uuid.uuid4().hex + ".tmp")
14
+ descriptor = os.open(temporary, os.O_WRONLY | os.O_CREAT | os.O_EXCL, 0o600)
15
+ try:
16
+ with os.fdopen(descriptor, "w") as stream:
17
+ json.dump(value, stream, indent=2, allow_nan=False)
18
+ stream.flush()
19
+ os.fsync(stream.fileno())
20
+ os.replace(temporary, path)
21
+ # Windows cannot open a directory to flush it; NTFS journals the rename itself.
22
+ if os.name != "nt":
23
+ parent = os.open(path.parent, os.O_RDONLY)
24
+ try:
25
+ os.fsync(parent)
26
+ finally:
27
+ os.close(parent)
28
+ finally:
29
+ temporary.unlink(missing_ok=True)
30
+
31
+
32
+ def fingerprint(value):
33
+ """SHA-256 of ``value`` as canonical JSON."""
34
+ return hashlib.sha256(
35
+ json.dumps(value, sort_keys=True, separators=(",", ":"), allow_nan=False).encode()
36
+ ).hexdigest()
37
+
38
+
39
+ def completion(path, operation_id):
40
+ """The completion receipt for ``operation_id`` at ``path``, or None before it exists."""
41
+ path = Path(path)
42
+ if path.is_symlink() or path.parent.is_symlink():
43
+ raise ValueError("Native receipt must not be a symbolic link")
44
+ if not path.exists():
45
+ return None
46
+ if not path.is_file() or not 1 <= path.stat().st_size <= 8 * 1024 * 1024:
47
+ raise ValueError("Native completion receipt is missing, unsafe or too large")
48
+ value = json.loads(path.read_text())
49
+ if (
50
+ not isinstance(value, dict)
51
+ or value.get("operation_id") != operation_id
52
+ or type(value.get("ok")) is not bool
53
+ ):
54
+ raise ValueError("Native completion receipt does not match the operation")
55
+ return value
56
+
57
+
58
+ def archive(path, record, resolution):
59
+ """Record how an interrupted operation was resolved; repeating the same record is a no-op."""
60
+ path = Path(path)
61
+ value = {"prior_operation": record, **resolution}
62
+ if path.exists():
63
+ if json.loads(path.read_text()) != value:
64
+ raise ValueError("Recovery record already exists with a different outcome")
65
+ return
66
+ write_json(path, value)