Add Copilot review skills, project instructions, and coding agent setup
for automated code review on regorus PRs.
Files added:
- .github/copilot-instructions.md — project context (no_std, 9 bindings,
dual execution paths, deny lints, security-critical evaluation)
- .github/skills/code-review/SKILL.md — fast single-agent review (~2 min)
- .github/skills/deep-review/SKILL.md — multi-agent deep review (~12 min)
- .github/copilot-setup-steps.yml — minimal coding agent environment
Development and testing methodology:
The skills were developed iteratively (v3 through v11.4) against a
460-line SARIF output module on the feature/sarif-output branch, which
served as a controlled test bed with 25 known issues of varying severity
(correctness, safety, API design, platform, security, performance).
Each version was tested by running the skill via the Copilot CLI, then
mapping discovered findings against the ground truth set to measure
recall and precision. Key iterations:
- v3: baseline single-agent (8/25 recall, 32%)
- v7: 3 parallel agents + verification (14/25, 56%)
- v10c: model diversity + adversarial pass (10/25, 40%)
- v11.3: merged adversarial-verifier architecture (12/25 + 2 novel, 0 noise)
- v11.4: domain expertise prompting (12/25 + 2 novel, 0 noise, full report)
The final architecture uses 3 parallel discovery agents (with cross-model
diversity and context asymmetry), risk-triggered micro-passes, and a
single adversarial verifier that both validates candidates via disproval
and hunts blind spots. Agents are prompted to reason from policy-author
perspective across Rego/OPA, Azure Policy, and RVM workloads.
Combined CR+DR catches 16-17/25 ground truth with zero false positives
and produces verified findings with confidence levels, test gap analysis,
and agent performance metrics.
Co-authored-by: Copilot <223556219+Copilot@users.noreply.github.com>
* build(deps): update all Rust dependencies to latest versions
Bulk-update all Cargo.lock files across the workspace and bindings
to their latest compatible versions. This supersedes the individual
per-directory dependabot PRs (#678-#682) that fail CI due to version
skew when only one lockfile is updated.
Co-authored-by: Copilot <223556219+Copilot@users.noreply.github.com>
* ci: refresh ALL Cargo lockfiles on dependabot PRs
Dependabot security updates bypass the grouped-updates config and
create per-directory PRs (one per Cargo.lock). This causes version
skew — e.g. rand gets bumped in bindings/ruby but stays old elsewhere,
breaking the build.
Fix by unconditionally refreshing all lockfiles whenever any Cargo
manifest or lockfile changes, rather than only the affected directory.
Also harden the workflow against expression injection:
- Move head.ref and base_ref to env vars (not inline ${{ }})
- Validate refs via git check-ref-format --branch
- Validate SHA format (hex, 40 chars) before use
- Fetch base branch by ref (not bare SHA) for reliable diffing
- Add security boundary comment on untrusted code checkout
- Add version comment on pinned checkout action SHA
Ref: https://github.com/dependabot/dependabot-core/issues/7547
Co-authored-by: Copilot <223556219+Copilot@users.noreply.github.com>
---------
Co-authored-by: Copilot <223556219+Copilot@users.noreply.github.com>
Update Cargo.lock to move rand to 0.10.1 so cargo-deny stops failing on RUSTSEC-2026-0097.
Also tighten the Python workflow cache boundaries by keying rust-cache to pinned runner images. The workflow now keeps Ubuntu 22.04, Ubuntu 24.04, and Windows 2022 caches separate, which avoids reusing host build artifacts across runner image changes. That is the class of issue behind the intermittent GLIBC mismatch seen in CI.
* build(csharp): rename NuGet package to Microsoft.Regorus
Align the NuGet package identity with the Microsoft.Regorus
namespace and the Microsoft.* reserved prefix on nuget.org
in preparation for publishing the package.
- Add explicit <PackageId>Microsoft.Regorus</PackageId>
- Update PackageVersion in Directory.Packages.props
- Update version bump regex for new package name
* build(csharp): default to ProjectReference for in-repo consumers
Use ProjectReference for Regorus.Tests, Benchmarks, TestApp,
and TargetExampleApp so that local development does not require
a pre-built .nupkg. PackageReference mode remains available via
/p:UsePackageReference=true for validating the packaged NuGet.
* build(csharp): add nuget.config with source mapping and xtask integration
Add explicit NuGet configuration to ensure in-repo builds always
resolve Microsoft.Regorus from the locally built package, even
after the package is published on nuget.org.
- Add nuget.config with <clear/> and packageSourceMapping
- Update xtask to copy .nupkg into local-packages/ directory
- Pass /p:UsePackageReference=true from xtask for CI testing
- Add restore validation step to the xtask test flow
- Add .gitignore for the local-packages directory
Without grouping, dependabot creates a separate PR per directory for the
same dependency. Each individual PR fails to build due to version skew
across the root workspace and binding crates.
Signed-off-by: Anand Krishnamoorthi <anakrish@microsoft.com>
* fix: update bindings and builtins for breaking dependency upgrades
- Update rand 0.10 API: use RngExt trait instead of removed Rng trait
- Update jsonschema 0.45 API: replace removed BasicOutput/apply with
iter_errors for schema validation
- Update jni 0.22 API: migrate from deprecated JNIEnv to EnvUnowned
with_env pattern, replace deprecated get_string/new_string/throw
methods with their modern equivalents
- Update pyo3 0.28 API: replace removed PyObject with Py<PyAny>,
deprecated downcast with cast, and removed with_gil with attach
Signed-off-by: Anand Krishnamoorthi <anakrish@microsoft.com>
* ci(dependabot): create per-dependency PRs for Cargo updates
Remove the groups.rust-dependencies catch-all group so Dependabot
opens a separate PR for each Cargo dependency update instead of
bundling them all into a single PR.
Signed-off-by: Anand Krishnamoorthi <anakrish@microsoft.com>
* fix: enable getrandom wasm_js feature for wasm32-unknown-unknown builds
Signed-off-by: Anand Krishnamoorthi <anakrish@microsoft.com>
* fix: address PR review comments
- Stream iter_errors directly into BTreeSet without intermediate Vec
- Use JNI_TRUE/JNI_FALSE for jboolean instead of bool coercion
Signed-off-by: Anand Krishnamoorthi <anakrish@microsoft.com>
---------
Signed-off-by: Anand Krishnamoorthi <anakrish@microsoft.com>
* build: consolidate dependabot cargo entries and add commit prefixes
Consolidate all 9 separate cargo ecosystem entries into a single entry
using the 'directories' key. This ensures Dependabot creates one PR per
dependency update across the root workspace and all bindings, preventing
version skew that caused build failures.
Also add semantic commit-message prefixes to all ecosystem entries:
- build(deps) for cargo, gomod, maven, nuget, pip, bundler
- ci(deps) for github-actions
Rename the cargo group to 'rust-dependencies' and the github-actions
group to 'github-actions' for clarity.
Signed-off-by: Anand Krishnamoorthi <anakrish@microsoft.com>
* fix: remove mimalloc from default features, fix indexmap/std propagation
Address #595: the vendored mimalloc allocator should not be imposed on
library consumers. Remove allocator-memory-limits and mimalloc from the
full-opa feature so that users of regorus as a library can choose their
own global allocator.
Bindings (ffi, java, python, ruby) that ship as standalone artifacts
continue to opt in to regorus/allocator-memory-limits explicitly so they
retain the performant allocator.
Also propagate indexmap/std via the std feature (using the indexmap?/std
weak-dependency syntax) so that users enabling std + rvm without default
features no longer hit 'IndexMap takes 3 generic arguments' errors.
Closes#595
Signed-off-by: Anand Krishnamoorthi <anakrish@microsoft.com>
* ci: add feature-combination checks to PR CI and weekly matrix
PR CI (xtask): add cargo check for 5 non-default feature combos in
run_ci_suite(). These run on every PR and catch compile failures from
feature-gating issues (e.g. #595) with near-zero overhead.
Weekly workflow: new feature-matrix.yml runs cargo build + cargo test
across 9 feature combinations every Monday. Uses a GitHub Actions matrix
with fail-fast: false so all combos are tested even if one fails.
Combinations tested weekly:
- std,arc (minimal library)
- std,arc,rvm (common library usage)
- std,arc,full-opa (full-opa without mimalloc)
- std,arc,full-opa,allocator-memory-limits (binding-style)
- std,arc,rvm,regex,time,semver,cache (cherry-picked builtins)
- std,arc,rvm,coverage,cache (observability)
- std,arc,full-opa,azure_policy (Azure Policy)
- std,arc,full-opa,azure-rbac (Azure RBAC)
- arc,opa-no-std (no_std codepath)
Signed-off-by: Anand Krishnamoorthi <anakrish@microsoft.com>
* fix: gate benchmark memory-limit calls behind allocator-memory-limits feature
The set_global_memory_limit function is only available when the
allocator-memory-limits feature is enabled. After removing mimalloc
from the default feature set, the rvm_benchmark failed to compile.
Add #[cfg(feature = "allocator-memory-limits")] guards around the
call sites and the MEMORY_LIMIT_BYTES constant.
Signed-off-by: Anand Krishnamoorthi <anakrish@microsoft.com>
---------
Signed-off-by: Anand Krishnamoorthi <anakrish@microsoft.com>
- Add cfg(not(miri)) guards to mimalloc module, global allocator, and
allocator-memory-limits code paths so Miri falls back to the default
system allocator instead of calling unsupported FFI functions.
- Set MIRIFLAGS="-Zmiri-disable-isolation" in the workflow so tests
that perform file I/O can run under Miri.
- Skip units/parse tests under Miri due to Float-vs-BigInt Number
representation mismatch with Miri's soft-float emulation.
Signed-off-by: Anand Krishnamoorthi <anakrish@microsoft.com>
Move RVM binary encoding from bincode to postcard and bump the format version. Update test helpers, docs, changelog, and refresh lockfiles after the swap.
Closes#575
Signed-off-by: Anand Krishnamoorthi <anakrish@microsoft.com>
- Expand dependabot coverage across Rust subcrates and other ecosystems.
- Group updates per dependency and ignore vendored mimalloc crates.
- Pin GitHub Actions to exact SHAs in existing workflows.
Additionally
- Include more metadata in nuget package
- Also generate snupkg for native symbols.
We intentionally don't add the symbols for native rust shared library
to the nuget package since that could increase the size of the nuget.
We will revisit that later.
- update licenses of all the bindings.
closes#551
Signed-off-by: Anand Krishnamoorthi <anakrish@microsoft.com>
- split the xtask crate into structured modules for
- bindings
- ci
- dev
- util
- no-std
- Adding commands for
- ci-release/ci-debug
- MUSL/no-std
- per- binding language smoke tests
- developer tasks (fmt, clippy, pre-commit, pre-push)
- refresh Cargo manifests/locks, binding readmes, and shared FFI helpers so every binding reuses the same preparation steps
- refactor GitHub Actions (release/debug, extensions, CodeQL, clippy, bindings) to call the new xtask commands
- Use rust-cache in ci workflows (microsoft qdk also does this)
- extend README with a contributor workflow section describing how xtask mirrors CI expectations
- update pre-commit and pre-push hooks to use the xtask dev commands
WORKAROUND:
When dotnet is run from an xtask, codeql tracer intercepts it an routes to a nonexistent binary.
Therefore in codeql workflow, xtask is not used for c# and instead dotnet is directly invoked.
Tracked by #545closes#475
Signed-off-by: Anand Krishnamoorthi <anakrish@microsoft.com>
Policy evaluation at scale needs to be able to set memory limits
so that a bad policy does not hog memory or to ensure that
policy evaluation itself does not use too much memory which could
cause other components to suffer.
This PR introduces capability to set and enforce global memory limits.
It also lays the groundwork for enabling per evaluation limits in future.
Once a global memory limit is set, Regorus maintains per thread counters
to track memory activity (allocation, deallocation) of a thread.
These counters are periodically flushed to global memory counters.
Per thread counters avoid the contention that updating global counters
on each alloc/free would cause.
Policy evaluation periodically checks these counters and raises errors
if allocated memory has exceeded the configured limit.
Currently memory limit capability is exposed only to FFI and C#.
Also update mimalloc to v2.2.6
Signed-off-by: Anand Krishnamoorthi <anakrish@microsoft.com>
This PR implements widely accepted Rust programming practices for
dealing with panics across ABI (programming language) boundaries.
- Add panic_guard.rs to wrap FFI calls and prevent panic across FFI/ABI boundary (undefined behavior).
- Capture per-thread backtraces via a temporary panic hook
- After a panic, subsequent invocations are poisoned.
- Integrate with_unwind_guard across the engine, schema registry, and target registry exportis
Signed-off-by: Anand Krishnamoorthi <anakrish@microsoft.com>
Lints are added (deny) at crate level.
In each offending file, the failing lints are explicitly allowed.
Each file will be fixed in subsequent PRs.
Signed-off-by: Anand Krishnamoorthi <anakrish@microsoft.com>
Add runtime detection for shared handle misuse
wrap the FFI engine handle with parking_lot::RwLock when the new
contention_checks feature is enabled, surfacing a clear “handle is already
in use” error instead of allowing undefined behavior
keep the feature optional so no_std builds or environments that supply
their own synchronization can opt out
caution users that this guards the handle itself but does not make the
engine’s operations globally thread-safe on its own
Signed-off-by: Anand Krishnamoorthi <anakrish@microsoft.com>
* feat: Add Azure RBAC condition parser
- declare an `azure-rbac` feature and expose the Azure RBAC module with parser, AST, and YAML-driven tests
- extend the shared lexer with RBAC-specific tokens, single-quoted strings, and corrected raw-string spans
- verify the parser via comprehensive test cases covering every operator and complex chaining
Signed-off-by: Anand Krishnamoorthi <anakrish@microsoft.com>
---------
Signed-off-by: Anand Krishnamoorthi <anakrish@microsoft.com>
* build: Add xtask automation for binding version management
Introduces a dedicated xtask crate that keeps language binding versions
in sync with the core regorus crate, following the workflow pattern used
by rust-analyzer, gitoxide, and ripgrep.
Key features:
- Git-based change detection: compares binding source files against a
base ref (merge-base with origin/main by default) plus unstaged/
untracked files to identify which bindings have been modified
- SemVer-aware bumping: binding edits trigger a minor version increment
(e.g. 0.5.1 → 0.6.0) under pre-1.0 semantics, signaling potential
breaking changes; clean bindings simply align to the root version
- Multi-language support: updates Cargo manifests (Rust FFI, Java,
Python, WASM, Ruby), Maven pom.xml (Java), Ruby version constants,
and C# project files in a single pass
- CI integration: --check mode fails fast when manifests are out of
sync, ensuring pre-commit and release-plz workflows catch stale
versions before merge
Integration points:
- release-plz.toml: runs cargo xtask bindings --base-ref origin/main
after bumping the root crate, so binding versions are updated
atomically during the release process
- scripts/pre-commit: invokes cargo xtask bindings --check to block
commits that would leave bindings out of sync
- .cargo/config.toml: defines cargo xtask alias for convenience
Documentation includes inline examples showing how version bumps behave
when bindings are ahead/behind the root, and notes that the minor
field acts as the major version under SemVer 0.y.z initial development
phase.
Signed-off-by: Anand Krishnamoorthi <anakrish@microsoft.com>
* build: refresh xtask tooling, workflows, and locks
- cargo xtask bindings: keep the binding version-sync pipeline intact
- cargo xtask update-deps: new helper to regenerate workspace/binding Cargo.lock files
- workflows: auto-detect the Java jar version in CI and temporarily disable the Ruby workflow
- lock files: refresh root + binding snapshots after the dependency sweep
Signed-off-by: Anand Krishnamoorthi <anakrish@microsoft.com>
---------
Signed-off-by: Anand Krishnamoorthi <anakrish@microsoft.com>
Security Improvements:
- Pin all GitHub Actions to specific commit hashes instead of version tags
- Update actions/checkout from v4 to commit 08eba0b27e820071cde6df949e0beb9ba4906955
- Update actions/setup-python from v5 to commit a26af69be951a213d495a4c3e4e4022e16d87065 (v5.6.0)
- Update actions/setup-java from v4 to commit dded0888837ed1f317902acf8a20df0ad188d165 (v5.0.0)
- Update actions/setup-node from v4 to commit 1e60f620b9541d16bece96c5465dc8ee9832be0b (v4.4.0)
- Update actions/setup-go from v5 to commit 41dfa10bad2bb2ae585af6ee5bb4d7d973ad74ed (v5.1.0)
- Update actions/setup-dotnet from v4 to commit 3e891b0cb619bf60e2c25674b222b8940e2c1c25 (v4.1.0)
- Update actions/upload-artifact from v4 to commit ea165f8d65b6e75b540449e92b4886f43607fa02 (v4.6.2)
- Update actions/download-artifact from v4 to commit 634f93cb2916e3fdff6788551b99b062d0335ce0 (v5.0.0)
- Update github/codeql-action from v3 to commit 01fe2e8c43536ad5e1085bad5e7cd6fbc8a30988 (v3.29.11)
Rust Toolchain Consolidation:
- Create custom composite action .github/actions/toolchains/rust/action.yml
- Standardize on Rust 1.89.0 (latest stable) with clippy and rustfmt components
- Add optional targets parameter for cross-compilation support
- Replace dtolnay/rust-toolchain@stable across 16 workflows
This creates a more secure, maintainable, and consistent CI/CD pipeline
with centralized Rust toolchain management across all workflows.
Signed-off-by: Anand Krishnamoorthi <anakrish@microsoft.com>
- Disable publishing for `ensure_no_std` test crate
- Add write content permissions to release-plz job expliclity
Signed-off-by: Burak Varlı <burakvar@amazon.co.uk>
Details:
- Implement complete Type enum with 12 variants: Any, Integer, Number, Boolean,
Null, String, Array, Set, Object, Enum, Const, AnyOf
- Add Schema wrapper struct with reference counting for efficient sharing
- Support JSON Schema-compatible deserialization with serde
- Implement discriminated subobjects for polymorphic type definitions
- Add comprehensive test suite covering all type variants
- Include Azure resource schema examples (Storage, VM, Key Vault, App Service)
- Create meta-schema validation system with lazy static validator
- Add extensive edge case and corner case test coverage
- Implement custom deserializers for complex schema patterns
This establishes the foundation for type checking and validation of Rego
policies, particularly useful for cloud resource schemas and policy validation.
Regorus's type system is a first of many features intended to
enable type checking and various other constraints on Rego policies.
The type system is inspired from:
- JSON schema
- Bicep
Signed-off-by: Anand Krishnamoorthi <anakrish@microsoft.com>
Also update binding versions and lock files.
Note:
- Ruby binding is not updated
- C# binding is v0.7.0. We will make it match Regorus version later.
Signed-off-by: Anand Krishnamoorthi <anakrish@microsoft.com>
- Specify compatibility = linux in pyproject.toml to ensure
manylinux compatibility.
- Use abi-py310 pyo3 feature to ensure compatibility with python
3.10 and later.
Signed-off-by: Anand Krishnamoorthi <anakrish@microsoft.com>
Organize C# binding example into separate Regorus nuget package and
a test app.
The nuget package targets netstandard 2.0 and 2.1.
The test app is tested for netframework 8.0.
Implement IDisposable for Engine cleanup.
Also remove net40 example. Can be added back later if needed.
Signed-off-by: Anand Krishnamoorthi <anakrish@microsoft.com>
Removed cryptographically insecure sha1. This existed only for OPA
compatibility.
Also exclude bindings from main workspace
Signed-off-by: Anand Krishnamoorthi <anakrish@microsoft.com>
Add `or` operator to Rego languages. Available via `rego-extensions`
Cargo feature.
If the evaluated lhs value is not false, null or undefined it is returned.
Otherwise rhs is evaluated and returned.
or operator has least precedence, and is left-associative.
closes#314
Also add test to lock down example policy path.
Also Fix clippy warning by using unwrap_or_default
Signed-off-by: Anand Krishnamoorthi <anakrish@microsoft.com>
