This release is a pre-release and may not be stable for production use.
Ragwell agent tools
Ask your agent questions about documents in your Ragwell project and get evidence you can check against the original sources. Ragwell retrieves relevant passages; your agent writes the explanation.
Local early access: the local adapter provides MCP search, bounded source expansion and a reusable Ragwell skill. Source reads passed guarded Codex and Claude Code tasks. Version 0.1.0a4 adds the owner's project description to authorized tool listings, helping agents recognize the connected knowledge. It requires published SDK 0.2.3 and a compatible API. Local use requires terminal setup, a ready Ragwell project and a dedicated API key. See the compatibility record for the earlier qualified clients and current delivery limits.
Hosted connection: use the production endpoint and OAuth client settings in the ChatGPT or Claude web guide. Earlier invited testers connected through ChatGPT web's Personal custom-MCP plugin flow or Claude web by signing in to Ragwell and approving one project. Linking, cited sample tasks/source reads and disconnect/reconnect passed. This route needs no local package or Ragwell API key. Managed workspaces and ordinary-user onboarding remain unqualified. See the compatibility record for actual test evidence; intended client support is not yet a certification.
| Your agent | Start here | Current route |
|---|---|---|
| Codex CLI | Connect Codex | Local MCP and optional skill |
| Claude Code | Connect Claude Code | Local MCP and optional skill |
| ChatGPT web / Personal | Connect ChatGPT | Hosted OAuth plugin; see qualification limits |
| Claude web / Free personal | Connect Claude web | Hosted OAuth connector; see qualification limits |
| Another MCP client | Generic setup | Qualify its local transport and credential handling |
Start with your agent guide above, or try the
four-document starter pack and
sample questions. Local setup is described in
the setup guide. You can ask how a process works,
compare policies, or consult decisions before making a change. The pilot cannot
upload/delete documents or switch projects. With document:read, the agent can
read additional context from a source returned by search, keeping its exact
document version and index generation.
For local connections, keys belong in agent credential settings, never chat. Searches consume Ragwell usage and returned text reaches your chosen agent provider. The API checks access on every search; read-only access can still disclose the permitted documents. Project descriptions also reach agents with current search access. After editing or clearing a description in Ragwell project settings, refresh the agent's tool list or reconnect if the client caches it.
For development, see CONTRIBUTING.md. The independent package uses the published Python SDK and requires no backend checkout. See how the tools work and tested compatibility for boundaries and supported routes.
Metadata
Release files for ragwell-agent-tools 0.1.0a4
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 | Actions |
|---|---|---|---|---|
| ragwell_agent_tools-0.1.0a4.tar.gz | 138.0 kB | With provenance | Details for ragwell_agent_tools-0.1.0a4.tar.gz |
Built distribution (wheel)
| File | Interpreter | ABI | Platform | Actions Reset |
|---|---|---|---|---|
| ragwell_agent_tools-0.1.0a4-py3-none-any.whl | Python 3 | none | any | Details for ragwell_agent_tools-0.1.0a4-py3-none-any.whl |
Total release size: 158.1 kB
Release files / ragwell_agent_tools-0.1.0a4.tar.gz
| Download URL | ragwell_agent_tools-0.1.0a4.tar.gz |
|---|---|
| Size | 138.0 kB |
| Tags | Source |
|
SHA-256 checksum How to use checksums |
323feb5bd0c0e7ba16fd8da5fd4c8fc2e7f51f23d1cf7ddda2e5acb12d58b3b6
|
|
BLAKE2b-256 checksum How to use checksums |
46191dc45085b7cb0eb065a2490c02cbf3fa3ff051fa716ba8db33e20b322d9d
|
| Upload date | |
|
Uploaded using Trusted Publishing? What is trusted publishing? |
Yes |
| Uploaded via |
twine/7.0.0 CPython/3.13.14
|
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 Oct 29, 2026.
Transparency logRelease files / ragwell_agent_tools-0.1.0a4-py3-none-any.whl
| Download URL | ragwell_agent_tools-0.1.0a4-py3-none-any.whl |
|---|---|
| Size | 20.0 kB |
| Tags | Python 3 |
|
SHA-256 checksum How to use checksums |
566324f2f3c07419022909da47c1dc27a5033d3a254a30f20fe7599e200f9572
|
|
BLAKE2b-256 checksum How to use checksums |
42f605776b4df4570d1ebc4002c524705e5fead13e5dcc14ca8e11cdfec0d721
|
| Upload date | |
|
Uploaded using Trusted Publishing? What is trusted publishing? |
Yes |
| Uploaded via |
twine/7.0.0 CPython/3.13.14
|
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 Oct 29, 2026.
Transparency logSecurity for ragwell-agent-tools 0.1.0a4
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 ragwell_agent_tools-0.1.0a4.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 Oct 29, 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.