MOOTx01 Memory Adapter
Drop-in /memories backend with governance for the Anthropic
memory_20250818 tool.
What it does
Any Claude 4+ / Fable 5 agent that thinks it is reading and writing
/memories files is actually reading and writing governed MOOTx01
drawers — with dedup, provenance, sensitivity floor, confirmation
state, and a full audit trail included.
The differentiator
Every model-written memory lands as unconfirmed with derived trust. A startup scan or audit pass can quarantine suspect lessons using the estate's contained filter and anomaly lenses. This is the poisoning defense the ecosystem currently lacks — the memory tool contract gives models write access; MOOTx01 gives you the governance to trust (or not trust) what they wrote.
Two surfaces
1. MCP tool (built into mootx01)
The mootx01 daemon registers a memory tool on its MCP surface that
matches Anthropic's contract exactly. Claude Code and Claude Desktop
users already connected to mootx01 as an MCP server get the memory tool
for free — no adapter needed.
2. Python SDK handler (this package)
For Messages API users who run the tool-use loop themselves:
from moot_memory import MootMemoryTool
import anthropic
client = anthropic.Anthropic()
memory = MootMemoryTool(base_url="http://127.0.0.1:4242")
runner = client.beta.messages.tool_runner(
model="claude-opus-4-8",
max_tokens=1024,
messages=[{"role": "user", "content": "Remember that Acme prefers email."}],
tools=[memory],
)
final = runner.until_done()
Or standalone without the SDK:
from moot_memory import MootMemoryHandler
handler = MootMemoryHandler(base_url="http://127.0.0.1:4242")
result = handler.execute({"command": "view", "path": "/memories"})
Op mapping
| memory_20250818 | Estate behavior |
|---|---|
| view (dir) | enumerate drawers in wing="memories" |
| view (file) | drawer content with line numbers |
| create | capture as unconfirmed drawer |
| str_replace | supersede: new content, old withdrawn (full lineage) |
| insert | supersede: content with insertion applied |
| delete | soft withdrawal (never hard-erase from model ops) |
| rename | capture at new location, withdraw old |
Security
- Path traversal protection: all paths validated against
/memories - Model writes land unconfirmed: poisoning quarantine seam
- Deletes are soft withdrawals: reversible, audit trail preserved
- Sensitivity floor: adapter wing is Normal
- File size cap: 100KB per file
- No hidden files: dotfiles rejected
Tests
# Against the running daemon
python -m pytest apps/moot-memory-adapter/tests/
# The poisoning quarantine test
python -m pytest apps/moot-memory-adapter/tests/test_memory_adapter.py::TestPoisoningQuarantine -v
Requirements
- Running mootx01 daemon (v1.0.22+)
- Python 3.10+
anthropicSDK (optional, for the BetaAbstractMemoryTool subclass)
1 maintainer
Metadata
1 maintainer
Release files for moot-memory 1.1.0
For a detailed explanation of source distributions (sdists) and built distributions (wheels), please see the package formats documentation.
Source distribution (sdist)
| File | Size | Uploaded | Provenance | |
|---|---|---|---|---|
| moot_memory-1.1.0.tar.gz | 7.1 kB | With provenance | Details |
Built distribution (wheel)
| File | Interpreter | ABI | Platform | Reset |
|---|---|---|---|---|
| moot_memory-1.1.0-py3-none-any.whl | Python 3 | none | any | Details |
Total release size: 12.6 kB
Release files / moot_memory-1.1.0.tar.gz
| Download URL | moot_memory-1.1.0.tar.gz |
|---|---|
| Size | 7.1 kB |
| Tags | Source |
|
SHA-256 checksum How to use checksums |
920c486d0b6ef8cd5d1fbc5c09e296a3cc1a918d0380239a6c5397f4841a7b1e
|
|
BLAKE2b-256 checksum How to use checksums |
99b6ed7e05328531f05bdaed5bac9fde142ce6165bb269e8109f35b2fcd6edad
|
| Upload date | |
|
Uploaded using Trusted Publishing? What is trusted publishing? |
Yes |
| Uploaded via |
twine/6.1.0 CPython/3.13.13
|
Provenance
Provenance describes where a file came from. On PyPI, provenance is shared via attestations, which provide a verifiable record of the build or publishing details. View details, limitations and caveats.
PyPI Publish Attestation
PyPI verified that this artifact, at this checksum, originated from the publisher listed below.
Signed by GitHub Actions, verified by PyPI on Sep 18, 2026.
Transparency logRelease files / moot_memory-1.1.0-py3-none-any.whl
| Download URL | moot_memory-1.1.0-py3-none-any.whl |
|---|---|
| Size | 5.5 kB |
| Tags | Python 3 |
|
SHA-256 checksum How to use checksums |
da1456066913a942f521df06faa14af209573aae87622793faebae90ab449c64
|
|
BLAKE2b-256 checksum How to use checksums |
2076d1dd021033893a3b96522b37e63023130832135d833c887652b63c960298
|
| Upload date | |
|
Uploaded using Trusted Publishing? What is trusted publishing? |
Yes |
| Uploaded via |
twine/6.1.0 CPython/3.13.13
|
Provenance
Provenance describes where a file came from. On PyPI, provenance is shared via attestations, which provide a verifiable record of the build or publishing details. View details, limitations and caveats.
PyPI Publish Attestation
PyPI verified that this artifact, at this checksum, originated from the publisher listed below.
Signed by GitHub Actions, verified by PyPI on Sep 18, 2026.
Transparency logSecurity for moot-memory 1.1.0
Provenance
Provenance describes where a file came from. On PyPI, provenance is shared via attestations, which provide a verifiable record of the build or publishing details. View details, limitations and caveats.
Full provenance
All 2 files in this release have been uploaded with provenance metadata.
Attestation for moot_memory-1.1.0.tar.gz
PyPI Publish Attestation
PyPI verified that this artifact, at this checksum, originated from the publisher listed below.
Signed by GitHub Actions, verified by PyPI on Sep 18, 2026.
Transparency logA note on package security
PyPI is a repository for distributing packages, not a service for auditing them. While third-party security researchers and our staff do their best to detect and remove malicious packages, PyPI cannot guarantee the quality, security or suitability of any given package.
To learn more about protecting yourself from malicious actors, visit our package security documentation.
