Metadata-Version: 2.5
Name: mylonite
Version: 0.8.6
Summary: Open-source AI-layer security testing — probes your AI agent for weaknesses and writes a validated regression test for each one it finds, gating CI.
Project-URL: Homepage, https://github.com/Abidemialade/mylonite
Project-URL: Source, https://github.com/Abidemialade/mylonite
Project-URL: Issues, https://github.com/Abidemialade/mylonite/issues
Project-URL: Changelog, https://github.com/Abidemialade/mylonite/blob/main/CHANGELOG.md
Author-email: Abidemi Alade <hello.mylonite@gmail.com>
Maintainer-email: Abidemi Alade <hello.mylonite@gmail.com>
License: 
                                         Apache License
                                   Version 2.0, January 2004
                                http://www.apache.org/licenses/
        
           TERMS AND CONDITIONS FOR USE, REPRODUCTION, AND DISTRIBUTION
        
           1. Definitions.
        
              "License" shall mean the terms and conditions for use, reproduction,
              and distribution as defined by Sections 1 through 9 of this document.
        
              "Licensor" shall mean the copyright owner or entity authorized by
              the copyright owner that is granting the License.
        
              "Legal Entity" shall mean the union of the acting entity and all
              other entities that control, are controlled by, or are under common
              control with that entity. For the purposes of this definition,
              "control" means (i) the power, direct or indirect, to cause the
              direction or management of such entity, whether by contract or
              otherwise, or (ii) ownership of fifty percent (50%) or more of the
              outstanding shares, or (iii) beneficial ownership of such entity.
        
              "You" (or "Your") shall mean an individual or Legal Entity
              exercising permissions granted by this License.
        
              "Source" form shall mean the preferred form for making modifications,
              including but not limited to software source code, documentation
              source, and configuration files.
        
              "Object" form shall mean any form resulting from mechanical
              transformation or translation of a Source form, including but
              not limited to compiled object code, generated documentation,
              and conversions to other media types.
        
              "Work" shall mean the work of authorship, whether in Source or
              Object form, made available under the License, as indicated by a
              copyright notice that is included in or attached to the work
              (an example is provided in the Appendix below).
        
              "Derivative Works" shall mean any work, whether in Source or Object
              form, that is based on (or derived from) the Work and for which the
              editorial revisions, annotations, elaborations, or other modifications
              represent, as a whole, an original work of authorship. For the purposes
              of this License, Derivative Works shall not include works that remain
              separable from, or merely link (or bind by name) to the interfaces of,
              the Work and Derivative Works thereof.
        
              "Contribution" shall mean any work of authorship, including
              the original version of the Work and any modifications or additions
              to that Work or Derivative Works thereof, that is intentionally
              submitted to Licensor for inclusion in the Work by the copyright owner
              or by an individual or Legal Entity authorized to submit on behalf of
              the copyright owner. For the purposes of this definition, "submitted"
              means any form of electronic, verbal, or written communication sent
              to the Licensor or its representatives, including but not limited to
              communication on electronic mailing lists, source code control systems,
              and issue tracking systems that are managed by, or on behalf of, the
              Licensor for the purpose of discussing and improving the Work, but
              excluding communication that is conspicuously marked or otherwise
              designated in writing by the copyright owner as "Not a Contribution."
        
              "Contributor" shall mean Licensor and any individual or Legal Entity
              on behalf of whom a Contribution has been received by Licensor and
              subsequently incorporated within the Work.
        
           2. Grant of Copyright License. Subject to the terms and conditions of
              this License, each Contributor hereby grants to You a perpetual,
              worldwide, non-exclusive, no-charge, royalty-free, irrevocable
              copyright license to reproduce, prepare Derivative Works of,
              publicly display, publicly perform, sublicense, and distribute the
              Work and such Derivative Works in Source or Object form.
        
           3. Grant of Patent License. Subject to the terms and conditions of
              this License, each Contributor hereby grants to You a perpetual,
              worldwide, non-exclusive, no-charge, royalty-free, irrevocable
              (except as stated in this section) patent license to make, have made,
              use, offer to sell, sell, import, and otherwise transfer the Work,
              where such license applies only to those patent claims licensable
              by such Contributor that are necessarily infringed by their
              Contribution(s) alone or by combination of their Contribution(s)
              with the Work to which such Contribution(s) was submitted. If You
              institute patent litigation against any entity (including a
              cross-claim or counterclaim in a lawsuit) alleging that the Work
              or a Contribution incorporated within the Work constitutes direct
              or contributory patent infringement, then any patent licenses
              granted to You under this License for that Work shall terminate
              as of the date such litigation is filed.
        
           4. Redistribution. You may reproduce and distribute copies of the
              Work or Derivative Works thereof in any medium, with or without
              modifications, and in Source or Object form, provided that You
              meet the following conditions:
        
              (a) You must give any other recipients of the Work or
                  Derivative Works a copy of this License; and
        
              (b) You must cause any modified files to carry prominent notices
                  stating that You changed the files; and
        
              (c) You must retain, in the Source form of any Derivative Works
                  that You distribute, all copyright, patent, trademark, and
                  attribution notices from the Source form of the Work,
                  excluding those notices that do not pertain to any part of
                  the Derivative Works; and
        
              (d) If the Work includes a "NOTICE" text file as part of its
                  distribution, then any Derivative Works that You distribute must
                  include a readable copy of the attribution notices contained
                  within such NOTICE file, excluding those notices that do not
                  pertain to any part of the Derivative Works, in at least one
                  of the following places: within a NOTICE text file distributed
                  as part of the Derivative Works; within the Source form or
                  documentation, if provided along with the Derivative Works; or,
                  within a display generated by the Derivative Works, if and
                  wherever such third-party notices normally appear. The contents
                  of the NOTICE file are for informational purposes only and
                  do not modify the License. You may add Your own attribution
                  notices within Derivative Works that You distribute, alongside
                  or as an addendum to the NOTICE text from the Work, provided
                  that such additional attribution notices cannot be construed
                  as modifying the License.
        
              You may add Your own copyright statement to Your modifications and
              may provide additional or different license terms and conditions
              for use, reproduction, or distribution of Your modifications, or
              for any such Derivative Works as a whole, provided Your use,
              reproduction, and distribution of the Work otherwise complies with
              the conditions stated in this License.
        
           5. Submission of Contributions. Unless You explicitly state otherwise,
              any Contribution intentionally submitted for inclusion in the Work
              by You to the Licensor shall be under the terms and conditions of
              this License, without any additional terms or conditions.
              Notwithstanding the above, nothing herein shall supersede or modify
              the terms of any separate license agreement you may have executed
              with Licensor regarding such Contributions.
        
           6. Trademarks. This License does not grant permission to use the trade
              names, trademarks, service marks, or product names of the Licensor,
              except as required for reasonable and customary use in describing the
              origin of the Work and reproducing the content of the NOTICE file.
        
           7. Disclaimer of Warranty. Unless required by applicable law or
              agreed to in writing, Licensor provides the Work (and each
              Contributor provides its Contributions) on an "AS IS" BASIS,
              WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or
              implied, including, without limitation, any warranties or conditions
              of TITLE, NON-INFRINGEMENT, MERCHANTABILITY, or FITNESS FOR A
              PARTICULAR PURPOSE. You are solely responsible for determining the
              appropriateness of using or redistributing the Work and assume any
              risks associated with Your exercise of permissions under this License.
        
           8. Limitation of Liability. In no event and under no legal theory,
              whether in tort (including negligence), contract, or otherwise,
              unless required by applicable law (such as deliberate and grossly
              negligent acts) or agreed to in writing, shall any Contributor be
              liable to You for damages, including any direct, indirect, special,
              incidental, or consequential damages of any character arising as a
              result of this License or out of the use or inability to use the
              Work (including but not limited to damages for loss of goodwill,
              work stoppage, computer failure or malfunction, or any and all
              other commercial damages or losses), even if such Contributor
              has been advised of the possibility of such damages.
        
           9. Accepting Warranty or Additional Liability. While redistributing
              the Work or Derivative Works thereof, You may choose to offer,
              and charge a fee for, acceptance of support, warranty, indemnity,
              or other liability obligations and/or rights consistent with this
              License. However, in accepting such obligations, You may act only
              on Your own behalf and on Your sole responsibility, not on behalf
              of any other Contributor, and only if You agree to indemnify,
              defend, and hold each Contributor harmless for any liability
              incurred by, or claims asserted against, such Contributor by reason
              of your accepting any such warranty or additional liability.
        
           END OF TERMS AND CONDITIONS
        
           APPENDIX: How to apply the Apache License to your work.
        
              To apply the Apache License to your work, attach the following
              boilerplate notice, with the fields enclosed by brackets "[]"
              replaced with your own identifying information. (Don't include
              the brackets!)  The text should be enclosed in the appropriate
              comment syntax for the file format. We also recommend that a
              file or class name and description of purpose be included on the
              same "printed page" as the copyright notice for easier
              identification within third-party archives.
        
           Copyright [yyyy] [name of copyright owner]
        
           Licensed under the Apache License, Version 2.0 (the "License");
           you may not use this file except in compliance with the License.
           You may obtain a copy of the License at
        
               http://www.apache.org/licenses/LICENSE-2.0
        
           Unless required by applicable law or agreed to in writing, software
           distributed under the License is distributed on an "AS IS" BASIS,
           WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
           See the License for the specific language governing permissions and
           limitations under the License.
License-File: LICENSE
License-File: NOTICE
Keywords: agentic-security,ai-security,llm-security,mcp,mitre-atlas,nist-ai-rmf,owasp,red-team,test-generation
Classifier: Development Status :: 4 - Beta
Classifier: Intended Audience :: Developers
Classifier: Intended Audience :: Information Technology
Classifier: License :: OSI Approved :: Apache Software License
Classifier: Operating System :: OS Independent
Classifier: Programming Language :: Python :: 3
Classifier: Programming Language :: Python :: 3.11
Classifier: Programming Language :: Python :: 3.12
Classifier: Programming Language :: Python :: 3.13
Classifier: Topic :: Security
Classifier: Topic :: Software Development :: Testing
Classifier: Typing :: Typed
Requires-Python: <3.14,>=3.11
Requires-Dist: aiohttp>=3.14.0
Requires-Dist: httpx>=0.27
Requires-Dist: json-repair>=0.30
Requires-Dist: jsonschema>=4.22
Requires-Dist: litellm>=1.83.10
Requires-Dist: mcp<2.0,>=1.0
Requires-Dist: pydantic-settings>=2.3
Requires-Dist: pydantic>=2.7
Requires-Dist: pytest>=8.2
Requires-Dist: python-dotenv>=1.2.2
Requires-Dist: pyyaml>=6.0
Requires-Dist: rich>=13.7
Requires-Dist: truststore>=0.9
Requires-Dist: typer>=0.12
Provides-Extra: dev
Requires-Dist: mypy>=1.10; extra == 'dev'
Requires-Dist: pre-commit>=3.7; extra == 'dev'
Requires-Dist: pytest-asyncio>=0.23; extra == 'dev'
Requires-Dist: pytest-cov>=5.0; extra == 'dev'
Requires-Dist: ruff==0.16.4; extra == 'dev'
Requires-Dist: types-jsonschema>=4.22; extra == 'dev'
Requires-Dist: types-pyyaml>=6.0; extra == 'dev'
Provides-Extra: docs
Requires-Dist: mkdocs-material>=9.5; extra == 'docs'
Requires-Dist: mkdocs>=1.6; extra == 'docs'
Provides-Extra: enterprise
Requires-Dist: truststore>=0.9; extra == 'enterprise'
Description-Content-Type: text/markdown

# Mylonite

> **Model robustness is not the same as application security.** A frontier model can
> resist every generic prompt-injection you throw at it and still hand an attacker a win,
> because the hole is in *your app's design*, not the model's alignment. Mylonite tests
> whether your **app-layer controls** are what stop the attack, writes a **validated
> regression test** for each weakness it finds, and gates CI so a model upgrade can't
> silently strip the protection away.

[![CI](https://github.com/Abidemialade/mylonite/actions/workflows/ci.yml/badge.svg)](https://github.com/Abidemialade/mylonite/actions/workflows/ci.yml)
[![PyPI](https://img.shields.io/pypi/v/mylonite.svg)](https://pypi.org/project/mylonite/)
[![GitHub release](https://img.shields.io/github/v/release/Abidemialade/mylonite)](https://github.com/Abidemialade/mylonite/releases/latest)
[![License](https://img.shields.io/badge/license-Apache--2.0-blue.svg)](./LICENSE)

**Built for teams shipping MCP or agentic apps who need CI-enforced regression coverage on
the AI layer.**

Point Mylonite at any MCP (Model Context Protocol) app, whatever model or framework is
behind it. It attacks the AI/agentic layer — the system prompt and tool/function schemas —
finds app-specific weaknesses, and for each one emits a
**validated, CI-gating `pytest` regression test**.

Mylonite deliberately does *not* test the surrounding traditional code; that work belongs to
SAST/DAST tools.

## The core idea

The differentiator is the **control-efficacy check**: hold the model constant, toggle only
the safeguard, and keep a finding only when the attack *fires* on your app and is *resisted*
once the control is applied — across a repeat-run filter that absorbs LLM randomness.

That answers a question a point-in-time scan cannot: **is your control carrying the
security, or is the model's current good behaviour carrying it?** Only one of those survives
a model upgrade.

Two fidelities, and Mylonite always tells you which one you got:

| Guarded side | What a KEPT verdict proves | How to get it |
|---|---|---|
| **Your own control**, toggled | Your implementation is load-bearing. The strongest result. | Declare `control_env` in your `target.yaml` |
| **A canonical control** applied at the adapter boundary | The attack is real and this class of control closes it — but the guarded side was Mylonite's shim, not your code | The default; works on any single-build app |

The second is genuinely useful and it is what runs on most real targets. It is not the same
claim as the first, and Mylonite will not print the stronger wording for it.

## Evidence

Mylonite maintains an [independent verification harness](./docs/verification.md) that scores
it against ground truth it did not author — runnable third-party MCP servers and published
academic benchmarks. **It publishes the misses as well as the hits.**

- **A KEPT external differential** on a third-party MCP email server: the attack fired
  **5/5** on the raw target, the guarded build leaked **0/5**, success-rate gap **1.00**.
- **Zero false positives** on a benign third-party server — the external precision baseline.
- **An external detection catch** on a peer-reviewed vulnerable MCP corpus.
- **0/8 recall** on one external challenge set, **0/60** on another flagged as vacuous, and
  an LLM-judge agreement F1 of **0.41**. Those are real results and they are published for
  the same reason the good ones are.

Every capability has been exercised against real third-party servers through the unchanged
CLI. Full scorecard, caveats included: [docs/verification.md](./docs/verification.md).

## A clean result is a result

Worth setting expectations before you run it. Against a well-designed app and a robust
model, Mylonite will often correctly find **nothing** — and that is the tool working, not
failing. A KEPT control-efficacy proof needs a weakness that actually lands, which in
practice means an app-design flaw (a consequential action with no approval step, an
unrestricted egress path) or an app configured to act autonomously.

This is why [Try it](#try-it) starts with the bundled reference app rather than your code:
it is deliberately vulnerable and finds weaknesses every time, so you can see the machinery
work before you point it somewhere the honest answer may be "nothing".

## Where this sits

Static scanners read your tool descriptions and flag the ones that look dangerous; you are
left to judge which flags matter. Model-eval harnesses swap models and score which one
behaves best. Mylonite does neither. It runs the attack against your app, then holds the
model constant and toggles only your safeguard — so the finding you get back is evidence
about *your control*, not about how a description reads or how a model scored today.

## Project status

**Beta, single maintainer.** As of v0.8.5 that is 241 commits from one contributor, with a
1,900-test suite and CI (ruff, mypy, pytest, pre-commit) enforced on every PR. The extension
contracts are versioned public API, but no third party has built a plugin against them yet.
If you are weighing this as a dependency in a security pipeline, pin a version — and read
[Known limitations](./docs/limitations.md) first.

## Install

```bash
pip install mylonite                    # the CLI, from PyPI
pip install mylonite mcp-kitchen-sink   # ...plus the bundled reference app
```

Python 3.11–3.13. (3.14 is not yet supported: `litellm` has no wheels for it.) Scanning
needs an LLM API key; `check`, `--scaffold` and `report` do not.

## Try it

Start with the bundled reference app. It is deliberately vulnerable, it runs in-process and
binds to nothing, and it is the fastest way to watch the differential actually fire — same
attacks, two builds, opposite results:

```bash
mylonite scan reference:vulnerable   # finds seeded weaknesses
mylonite scan reference:guarded      # same attacks, comes up clean
```

That contrast *is* the product. See [the reference app](./docs/quarry.md) for what is seeded
in it and why. (Both commands call a model, so they need an API key.)

### Then point it at your own app

**The first two steps are free** — no API key, no model call, no spend.

```bash
# 1. Introspect a server and write a starter target.yaml
mylonite scan --command "python" --arg "my_server.py" --scaffold app.yaml --scope my-app

# 2. Static structural pre-check of that tool surface
mylonite check --target-file app.yaml
```

`--scaffold` connects to your server, lists its tools, tells you which weakness classes
apply to that surface, and flags the consequential-action tools worth guarding. Treat it as
a scope check, not a verdict: it reads your tool *surface*, not your tool descriptions, and
everything it suggests is a hint for you to confirm.

`check` reports structural exposure — consequential tools with no approval step,
descriptions that steer the agent, tools taking a network destination, unpinned
descriptions. `--enforce` turns it into a CI gate: it exits non-zero on the substantive
W1–W4 structural findings and treats the "unpinned descriptions" advisory (which fires on
every tool of every server on first contact) as a suggestion, not a gate — so it is
adoptable on day one.

Proving which weaknesses actually *land*, and which of your controls stops them, is the scan
itself. That needs a key:

```bash
mylonite scan --target-file app.yaml --authorize my-app
```

Expect this to find less than the reference app did — often nothing. See
[A clean result is a result](#a-clean-result-is-a-result) above, and
[docs/limitations.md](./docs/limitations.md) for where the tool's reach genuinely ends.

## From scan to a gating PR

`mylonite gate` runs the whole pipeline — find an exploit, write a regression test, validate
it against the control-efficacy check, and optionally open a PR that gates CI on it:

```bash
mylonite gate reference:vulnerable                                   # find -> test -> validate
mylonite gate --target-file app.yaml --authorize my-app --open-pr    # ...and open the PR
```

**`gate` does not touch your repository unless you ask it to.** By default it writes its
artefacts under `.mylonite/gate/` — the regression test, the exploit JSON, the validation
report — and prints the exact `git` and `gh` commands to commit and open the PR yourself.
Add `--open-pr` to have it create the branch, commit, and open the PR; add `--workflows` to
scaffold the two CI templates (a cheap per-PR gate and a nightly discovery run).

The PR carries the finding, its OWASP/ASI/ATLAS/NIST tags, the validation evidence, and an
evidence-anchored recommended fix naming the actual tool and argument that landed the
exploit. Full guide: [docs/ci-gating.md](./docs/ci-gating.md). Behind a corporate network,
see [docs/enterprise-networking.md](./docs/enterprise-networking.md).

## Commands

| Command | What it does | Needs a key? |
|---|---|---|
| `mylonite check` | Static structural pre-check of a tool surface. `--enforce` makes it a CI gate. | No |
| `mylonite scan` | The exploit-finding loop. `--scaffold` introspects a server and writes a starter `target.yaml`. | Yes (except `--scaffold`) |
| `mylonite generate` | Emits the `pytest` regression test from a confirmed exploit. | No |
| `mylonite validate` | Proves an emitted test is meaningful via the control-efficacy check. `--fast` skips it for a weaker gate. | Yes |
| `mylonite gate` | The end-to-end flow: scan → generate → validate → optionally open a gating PR. | Yes |
| `mylonite ablate` | Scores each safeguard as load-bearing, redundant, or security theater. Needs a target file. | Yes |
| `mylonite report` | Terminal trust panel, **SARIF 2.1.0**, or a JSON bundle — all carrying the differential proof and compliance tags. | No |
| `mylonite plugins` | Lists installed extension plugins across all five contract groups. | No |
| `mylonite version` | Prints the installed version. | No |

Exit codes are a documented contract (`0` success · `1` findings · `2` config · `3` budget ·
`4` provider · `5` not kept · `6` generate failed · `7` validate failed · `8` gate PR step
failed). Full details in the [CLI reference](./docs/cli-reference.md).

Remote MCP transport (SSE / streamable-HTTP), versioned extension contracts, and
entry-point plugins are covered in the [architecture guide](./docs/architecture.md).

## Compliance metadata

Every emitted test and every finding carries tags from four frameworks: **OWASP LLM Top 10
2025**, **OWASP ASI 2026**, **MITRE ATLAS**, and **NIST AI RMF**. They ride into the pytest
markers, the SARIF output and the JSON bundle, so a finding is traceable to the control
catalogue your auditors already use. See [docs/standards-mapping.md](./docs/standards-mapping.md).

## Documentation

**Full docs site:** [abidemialade.github.io/mylonite](https://abidemialade.github.io/mylonite/)
(or `mkdocs serve` from a checkout). Highlights:

- [Quickstart](./docs/quickstart.md) · [Test your own app](./docs/test-your-app.md) — install and point it at your MCP server.
- [Weakness classes](./docs/weakness-classes.md) · [Attack modes](./docs/attack-modes.md) — what's tested and how attacks work.
- [The validation engine](./docs/validation.md) — the control-efficacy check and the differential.
- [Independent verification](./docs/verification.md) — the honest scorecard against ground truth Mylonite didn't author.
- [Known limitations](./docs/limitations.md) — where the tool's reach ends, in one place.
- [Reading the results](./docs/reading-results.md) · [CLI reference](./docs/cli-reference.md) · [target.yaml](./docs/target-file.md).
- [CI gating](./docs/ci-gating.md) · [Architecture](./docs/architecture.md) · [Plugin authoring](./docs/plugin-authoring.md).
- [ROADMAP.md](./ROADMAP.md) · [CONTRIBUTING.md](./CONTRIBUTING.md) · [GOVERNANCE.md](./GOVERNANCE.md) · [SECURITY.md](./SECURITY.md).

## Responsible use

Mylonite reproduces working weaknesses in AI agents. **Use it only against targets you
control or are contractually authorized to test.** Every command that live-drives a real
target — `scan`, `gate`, `validate`, and `ablate` — refuses to run without an explicit
`--authorize` flag naming that target: the value must equal the target's declared `scope`,
or its family name when no scope is declared. The bundled vulnerable reference agent runs
in-process and binds to nothing.

Full policy: [SECURITY.md](./SECURITY.md).

## Contributing

Bug reports, adapter requests, and attack-pattern submissions are welcome — see
[CONTRIBUTING.md](./CONTRIBUTING.md) for dev setup, how to author a plugin, and the PR
conventions. The five extension points (attack modules, test generators, validators, target
adapters, compliance mappers) are versioned public API with reference implementations
in-repo.

## License

Apache License 2.0. See [LICENSE](./LICENSE) and [NOTICE](./NOTICE).
