horustrace 0.4.0__tar.gz

horustrace-0.4.0/LICENSE

@@ -0,0 +1,202 @@
+
+                                 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.

horustrace-0.4.0/PKG-INFO

@@ -0,0 +1,430 @@
+Metadata-Version: 2.4
+Name: horustrace
+Version: 0.4.0
+Summary: Policy-as-code security analysis for AI agents
+Author: HorusTrace contributors
+License: Apache-2.0
+Project-URL: Homepage, https://github.com/psandhir/horustrace
+Project-URL: Repository, https://github.com/psandhir/horustrace
+Project-URL: Issues, https://github.com/psandhir/horustrace/issues
+Project-URL: Security, https://github.com/psandhir/horustrace/security/policy
+Keywords: ai-security,agents,mcp,openai,google-adk,gcp,security,static-analysis
+Classifier: Development Status :: 4 - Beta
+Classifier: Environment :: Console
+Classifier: Intended Audience :: Developers
+Classifier: Intended Audience :: Information Technology
+Classifier: License :: OSI Approved :: Apache Software License
+Classifier: Programming Language :: Python :: 3
+Classifier: Programming Language :: Python :: 3.11
+Classifier: Programming Language :: Python :: 3.12
+Classifier: Topic :: Security
+Requires-Python: >=3.11
+Description-Content-Type: text/markdown
+License-File: LICENSE
+Requires-Dist: PyYAML<7,>=6.0.2
+Provides-Extra: dev
+Requires-Dist: pytest<9,>=8.3; extra == "dev"
+Requires-Dist: ruff<1,>=0.12; extra == "dev"
+Dynamic: license-file
+
+# HorusTrace
+
+[![CI](https://github.com/psandhir/horustrace/actions/workflows/ci.yml/badge.svg)](https://github.com/psandhir/horustrace/actions/workflows/ci.yml)
+[![License](https://img.shields.io/badge/license-Apache--2.0-blue.svg)](LICENSE)
+[![Python](https://img.shields.io/badge/python-3.11%2B-blue.svg)](pyproject.toml)
+
+**Five-layer policy-as-code security analysis for AI agents.**
+
+HorusTrace statically discovers agent configuration and evaluates five connected security layers:
+
+1. **Agent configuration** — tools, approvals, guardrails, MCP, code execution and framework-specific controls.
+2. **Capability analysis** — effective authority, capability budgets, prohibited actions and dangerous combinations.
+3. **Identity & permissions** — cloud/IAM roles, wildcard permissions, OAuth scopes and credential source.
+4. **Data & network reachability** — sensitive resources, resource scope, outbound destinations and allowlist violations.
+5. **Attack-path analysis** — potential risk combinations such as untrusted content → delegated agent → shell, or confidential data → agent → external write.
+
+> Status: **v0.4 development** on the `v0.4-dev` branch; this branch is being prepared for the first HorusTrace v0.4.0 release. Findings are deterministic within supported constructs. HorusTrace does not prove runtime exploitability or complete live cloud authority.
+
+## Security model
+
+```text
+What is configured?
+        ↓
+What can the agent do directly or through delegation?
+        ↓
+What authority does its identity have?
+        ↓
+What data/resources/destinations are reachable?
+        ↓
+Which end-to-end attack paths exist?
+```
+
+The scanner is **static-first and local-first**. It does not import target Python modules and does not launch MCP servers or agents.
+
+## Current framework/input coverage
+
+- **Google Agent Development Kit (ADK) Python 2.x — first-class adapter**.
+- **Google ADK Agent Config YAML** (`root_agent.yaml` and related agent configs).
+- OpenAI Agents SDK Python constructs, including v0.4 handoff normalization.
+- Initial LangGraph `StateGraph` / `MessageGraph` normalization in v0.4 development.
+- Common MCP JSON configuration (`mcp.json`, `.mcp.json`).
+- Framework-neutral `horustrace.manifest.yaml` for business/security intent.
+- Terraform (`.tf`) for an initial GCP/Azure/AWS IAM view.
+- ADK `.env` credential-source checks without exposing secret values in findings.
+
+## Google ADK coverage
+
+The v0.3 release (published as HorusTrace) performs repository-aware analysis of security-relevant ADK composition rather than only matching individual `Agent(...)` declarations.
+
+### Agents and orchestration
+
+- `Agent` / `LlmAgent`.
+- `SequentialAgent`, `ParallelAgent`, `LoopAgent`.
+- `sub_agents` and transitive delegated authority.
+- `AgentTool` delegation and `include_plugins` isolation.
+- `RemoteA2aAgent` consumption and `to_a2a(...)` exposure.
+- transfer restrictions and workflow metadata.
+- cross-file Python/config delegation aliases.
+
+### Tools and execution
+
+- plain Python functions used as ADK tools.
+- `FunctionTool`, `LongRunningFunctionTool`, `AuthenticatedFunctionTool`.
+- `require_confirmation`.
+- `ExecuteBashTool` and `BashToolPolicy`.
+- `EnvironmentToolset` / `LocalEnvironment`.
+- `UnsafeLocalCodeExecutor`, `BuiltInCodeExecutor`, `AgentEngineSandboxCodeExecutor`, `GkeCodeExecutor`.
+- `ComputerUseToolset`.
+- BigQuery/Bigtable/Data Agent toolsets.
+- Google API/Gmail/Calendar/Docs/Sheets/Slides/YouTube toolsets.
+- search/retrieval tools and inferred untrusted external content.
+- OpenAPI/API Hub/Application Integration/REST-style tool surfaces.
+- memory/artifact/MCP-resource tools represented as data capabilities.
+
+### MCP
+
+- `McpToolset` / `MCPToolset`.
+- stdio, SSE and Streamable HTTP connection parameters.
+- URL/transport analysis.
+- auth scheme/credential/header-provider detection.
+- recognized authorization headers.
+- tool filters.
+- confirmation requirements.
+
+### Safety controls
+
+- `before_agent_callback`, `after_agent_callback`.
+- `before_model_callback`, `after_model_callback`.
+- `before_tool_callback`, `after_tool_callback`.
+- model/tool error callbacks.
+- security/safety/policy-style plugins attached through `App`/`Runner`.
+- tool-level confirmation.
+
+### Google identity context
+
+- `google.auth.default(scopes=...)`.
+- service-account-file use.
+- ADK credentials config objects.
+- Google API toolset `additional_scopes`.
+- client-secret literals (reported without secret material).
+- `.env` API-key/service-account-file indicators.
+- Terraform GCP IAM bindings feeding Layer 3.
+
+See [`docs/google-adk.md`](docs/google-adk.md) for the exact supported surface and limitations.
+
+## Quick start
+
+
+```bash
+python -m venv .venv
+source .venv/bin/activate
+pip install horustrace
+horustrace scan .
+```
+
+List the built-in rule catalogue without scanning a project:
+
+```bash
+horustrace rules
+horustrace rules --format json --output rules.json
+```
+
+### ADK demo
+
+```bash
+horustrace scan examples/google-adk-vulnerable --fail-on none
+horustrace scan examples/google-adk-secure --fail-on none
+```
+
+The vulnerable ADK fixture intentionally exercises all five layers. The secure fixture should return zero findings under the current rule catalogue.
+
+Generate SARIF:
+
+```bash
+horustrace scan . --format sarif --output horustrace.sarif --fail-on none
+```
+
+Fail CI on high/critical findings:
+
+```bash
+horustrace scan . --fail-on high
+```
+
+Malformed, unreadable, or structurally invalid policy manifests stop the scan with exit
+code 1 and an error on stderr. This also applies with `--fail-on none`; that option
+only disables failure for security findings. No report is written for a failed scan.
+
+### Coverage and strict CI
+
+```bash
+horustrace scan . --strict --format json --output report.json
+```
+
+All report formats include file counts and coverage diagnostics. JSON exposes a
+`coverage` object; SARIF exposes coverage in run properties and diagnostics as tool
+execution notifications. Console output separates coverage from findings by layer.
+`--strict` returns exit code 1 for detected incomplete analysis, regardless of
+`--fail-on`. Unlike a fatal manifest error, incomplete analysis still writes the
+report so CI can retain the diagnostics. Security threshold failures return exit
+code 2; incomplete analysis takes precedence in strict mode.
+
+Diagnostics use stable `ARG-COV-*` identifiers and currently cover read/parse
+failures, unresolved Python tool/MCP references, unresolved delegation, dynamic
+agent configuration sequences or expanded keyword arguments, unresolved external
+helper semantics, dynamic MCP endpoints/tool filters, unknown MCP authentication
+state, and scans with no supported security targets.
+Files in default ignored directories, and subtrees containing an
+`.horustrace-ignore` marker, are excluded from file counts. Other unsupported
+file types are counted as skipped. A scanned file was read and parsed;
+that count does not mean its entire application behavior was understood.
+
+No detected coverage gaps is not proof of complete analysis. Runtime-generated
+behavior, arbitrary function semantics, cloud authorization, and control effectiveness
+remain outside these diagnostics. A clean report means no supported rules triggered.
+
+### Fingerprints, baselines and suppressions
+
+Every finding has an `arg-v1:` fingerprint based on its rule, agent, repository-relative
+path, and evidence. Line numbers and checkout roots are excluded, so fingerprints
+survive routine source movement and different CI workspaces. A material evidence or
+scope change produces a new fingerprint.
+
+Create an initial baseline of current findings:
+
+```bash
+horustrace baseline . \
+  --output .horustrace.suppressions.yaml \
+  --reason "Initial adoption backlog SEC-42" \
+  --expires 2026-12-31
+```
+
+Baseline creation scans without applying existing suppressions. It requires a reason
+and a non-past expiry and refuses to replace an existing file unless `--force` is
+passed. It also refuses to write when coverage diagnostics show incomplete analysis.
+Review the generated entries before committing them; a baseline records
+temporary risk acceptance rather than making the findings safe.
+
+Suppression files use this schema:
+
+```yaml
+version: 1
+suppressions:
+  - id: accepted-shell-migration
+    reason: Temporary migration path owned by SEC-42
+    expires: 2026-12-31
+    fingerprint: arg-v1:0123456789abcdef01234567
+    rule_id: AGT020
+```
+
+A fingerprint is the narrowest scope. A rule suppression without a fingerprint must
+also specify `agent` or a repository-relative `path` glob. IDs, reasons, and expiry
+dates are mandatory; unknown fields and duplicate IDs or YAML keys fail closed.
+IDs use letters, digits, dots, underscores, and dashes; reasons are single-line;
+path scopes cannot be absolute or escape the scan root.
+Expired entries never hide findings. Stale, matched, and expired entries remain in
+console, JSON, and SARIF audit output. `--strict` also fails when an exception has
+expired. Use `--suppressions path/to/file.yaml` to select a non-default file.
+
+### Reviewed benchmark
+
+```bash
+horustrace benchmark benchmarks/cases.yaml
+```
+
+The reviewed corpus declares the exact `RULE@agent` findings expected for each case.
+Unexpected findings are measured as false positives and missing findings as false
+negatives. The v0.3 corpus contains 26 reviewed scenarios across secure, execution,
+delegation, MCP, identity, data/network, attack-path and dynamic/unresolved analysis.
+One case intentionally expects incomplete analysis and an exact coverage diagnostic;
+all other cases fail on incomplete coverage. The command exits nonzero on any drift
+and supports `--format json` for CI artifacts. See
+[`benchmarks/README.md`](benchmarks/README.md).
+
+### Evidence and control semantics
+
+Findings retain their rule IDs and severity thresholds and now include:
+
+- `assessment`: `static_configuration`, `policy_violation`, `heuristic_risk`, or
+  `potential_risk`.
+- `provenance`: facts with a subject, origin, and source location. `observed` means
+  a supported static configuration was discovered; `declared` means a manifest
+  assertion; `inferred` means a heuristic or derived relationship.
+- `limitations`: uncertainty about runtime authority, controls, and exploitability.
+
+Configuration findings include local evidence; aggregate capability/data/path
+findings include agent context, which is not a formal data-flow trace. A manifest
+policy violation may involve declared or inferred capabilities; it does not establish
+runtime authority. Function capabilities are heuristic, while recognized built-in
+capabilities follow the scanner's supported static semantics.
+
+`control_observations` in JSON and SARIF distinguish approval configuration,
+callback hooks, plugin-name inference, sandbox configuration, and possible network
+destinations. For supported ADK and MCP configuration, they also record static
+evidence of sandbox timeout/network/filesystem limits, Bash allowlist plus blocklist
+policies, MCP tool allowlists, and whether every discovered egress destination fits
+a declared allowlist. Their runtime effectiveness is always `not_verified`.
+Callback or plugin presence can satisfy a missing-hook rule, but does not prove that
+arbitrary callback/plugin code authorizes actions safely. Approval callbacks alone
+do not establish an approval requirement. Mixed hosted-MCP approval policies are
+treated as unknown rather than blanket approval.
+
+Attack paths are labeled potential risks with `basis: capability_cooccurrence` and
+`exploitability: not_verified`. Their severity reflects potential impact, not proven
+exploitability. A literal URL inside a function establishes a possible destination,
+not an egress allowlist; such functions now trigger the missing-restriction rule.
+No runtime enforcement is inferred from those literals.
+
+### Manifest schema
+
+Manifests support schema version `1`. Omitting `version` retains the legacy v1
+behavior. Unknown fields, unsupported versions, duplicate YAML keys, invalid nested
+types, cyclic YAML aliases, and negative capability limits are rejected. Approval,
+guardrail, authentication, and restriction fields require actual booleans;
+capability/scope fields accept strings or lists of strings. Policy errors identify
+the field and source line/column without printing its value. Existing documented
+field aliases remain supported. The schema definitions live in
+[`manifest_schema.py`](src/horustrace/manifest_schema.py).
+
+
+## ADK-specific rule highlights
+
+- `ADK001` — privileged ADK agent lacks a detected tool-control callback/plugin/confirmation boundary.
+- `ADK002` — unsafe local code executor.
+- `ADK003` — `LocalEnvironment` exposes local shell/file I/O.
+- `ADK004` — bash execution without a detected restrictive `BashToolPolicy`.
+- `ADK012` — sandboxed code execution lacks an explicit timeout, network, or filesystem limit.
+- `ADK005` — computer-use capability lacks an explicit action boundary.
+- `ADK006` — BigQuery write capability is not statically blocked.
+- `ADK007` — broad generated/API toolset without a tool filter.
+- `ADK008` — delegated `AgentTool` disables inherited plugins.
+- `ADK009` — remote A2A agent card uses plaintext HTTP.
+- `ADK010` — remote A2A agent has no detected authentication.
+- `ADK011` — privileged agent is exposed over A2A without a detected safety control.
+
+These run in addition to the framework-neutral AGT/CAP/IDN/DATA/NET/PATH rules.
+
+## Framework-neutral security manifest
+
+The manifest declares business intent and runtime context that static source parsing cannot prove:
+
+```yaml
+version: 1
+agents:
+  - name: invoice-agent
+
+    inputs:
+      - name: supplier-portal
+        kind: web
+        trust: untrusted
+
+    data:
+      - name: invoices
+        classification: confidential
+        selector: /finance/invoices/**
+
+    identities:
+      - name: invoice-agent-sa
+        provider: gcp
+        roles: [roles/storage.objectViewer]
+        resource_scope: projects/acme/buckets/invoices
+        credential_source: workload_identity
+
+    network:
+      - target: https://erp.example.com/**
+        restricted: true
+
+    policy:
+      required: [data.read, external.write, network.external]
+      denied_capabilities: [process.execute, destructive.write]
+      allowed_resources: [/finance/invoices/**]
+      allowed_destinations: [https://erp.example.com/**]
+      require_approval_for: [external.write]
+      max_privileged_capabilities: 1
+```
+
+This enables least-privilege comparison between **required** and **effective** capabilities and lets Layers 4–5 reason about data and network paths.
+
+## GitHub Action
+
+```yaml
+- uses: psandhir/horustrace@v0.4.0
+  with:
+    path: .
+    fail-on: high
+    strict: "true"
+    suppressions: .horustrace.suppressions.yaml
+```
+
+## Design principles
+
+1. **Do not execute the target.** Static analysis must be safe on untrusted repositories.
+2. **Separate observation from policy.** Adapters discover facts; policy adds business/security intent.
+3. **Analyse effective authority.** Direct and delegated tool combinations matter more than isolated calls.
+4. **Connect identity, data and egress.** Agent risk is an end-to-end property.
+5. **Explain the path.** Findings include nodes forming the attack chain.
+6. **Prefer deterministic CI findings.** Semantic/LLM analysis can be additive later.
+
+## Scope boundary
+
+The Google ADK adapter is intended to be comprehensive for **security-relevant static constructs in current Python ADK 2.x and native Agent Config YAML**. It is not a claim that arbitrary third-party tool implementations, dynamically generated Python, runtime cloud authorization, or separate Java/Go/JavaScript/Kotlin ADK SDK syntax is fully analysed. Those require dedicated adapters or runtime/cloud-control-plane enrichment.
+
+HorusTrace is not a runtime firewall, formal taint verifier, malware scanner or proof that a prompt injection is exploitable. It does not execute the application or call cloud control planes during a normal scan.
+
+## Project roadmap
+
+See [ROADMAP.md](ROADMAP.md) for planned live GCP authority resolution, change-aware analysis, reachability enrichment, and framework expansion.
+
+## Support
+
+See [SUPPORT.md](SUPPORT.md).
+
+## Security
+
+See [SECURITY.md](SECURITY.md).
+
+## Contributing
+
+See [CONTRIBUTING.md](CONTRIBUTING.md).
+
+## License
+
+Apache-2.0. See [LICENSE](LICENSE).
+
+### Repository scanner configuration
+
+Use `.horustrace.yaml` to tune scanner policy separately from the security intent manifest and temporary suppressions:
+
+```yaml
+version: 1
+scanner:
+  strict: true
+rules:
+  ADK007: {severity: high}
+  AGT022: {enabled: false}
+```
+
+Use `horustrace scan . --config path/to/config.yaml` to select a file explicitly. Disabled rules are reported separately from suppressions; severity overrides affect reporting and failure thresholds but not rule metadata or finding fingerprints.
+
+### v0.3 release notes
+
+v0.3 (released as HorusTrace) moved the project from primarily file-level ADK parsing toward repository-level security reachability analysis. It adds cross-file tool and helper resolution, conservative factory and collection resolution, static MCP constant resolution, improved ADK execution/control semantics, stronger identity and OAuth linkage, and lower-noise network and capability inference. Coverage gaps remain explicit rather than being treated as safe. See [`docs/releases/v0.3.0.md`](docs/releases/v0.3.0.md) for the release summary.