starkware-libs
- 21 skills
- 0 followers
- 10 hours ago last updated
- ▌ Concise Comments · starkware-libsConcise comments and documentation style for code comments, docstrings, and module/file docs. Use when writing, editing, or reviewing comments, doc-comments, struct/field docs, constants, or entry (lib/index) files. Encodes Yoni-Starkware's review preferences — concise, present-tense, no history or design-doc references.
- ▌
- ▌ Pr Splitter · starkware-libs bundleSplit a large work-in-progress PR into a stack of small, reviewable PRs. Use this skill whenever the user mentions splitting a PR, stacking PRs, breaking up a large diff, making a PR easier to review, creating a PR stack, or decomposing changes into incremental commits. Also trigger when the user says things like "this PR is too big", "reviewers are struggling with this", or "I need to break this into smaller pieces". Works with any git-based project and any language.
- ▌ Debugging Wizard · starkware-libs bundleUse when investigating errors, analyzing stack traces, or finding root causes of unexpected behavior. Invoke for error investigation, troubleshooting, log analysis, root cause analysis.
- ▌
- ▌ Orizi Review · starkware-libsCode review in the style of Ori Ziv (orizi). Reviews for correctness, simplicity, clean code, and performance. Catches unnecessary complexity, redundant code, naming issues, and architectural concerns. Use /orizi-review [file or PR] to review changed code.
- ▌ Writing Tests · starkware-libs bundleUse when writing or changing tests in this repo, deciding what to test or where a test belongs, adding a regression test for a bug fix, reviewing the test portion of a PR, deciding which golden outputs to bless and reviewing the blessed diff, or adding a benchmark. Trigger phrases include "add a test", "write tests for", "is this covered", "does this need a test", "review the tests here", "which layer should this test go in". Also use for the mechanics of the golden framework itself — the //! > file format, creating test_file_test! runners, or how CAIRO_FIX_TESTS works (covered in references/golden-framework.md).
- ▌ Triage CI Failure · starkware-libs bundleTriage CI failures, flaky tests, and broken builds in the sequencer mono-repo. Use when a user mentions a failing CI job, flaky test, red check, or shares a GitHub Actions / PR URL — the skill pulls failure context directly from the public GitHub REST API so you can usually diagnose and report a verdict without asking the user any follow-up questions.
- ▌ Run Network Benchmark · starkware-libsRun the apollo_network_benchmark P2P stress test end-to-end — gather parameters, start a local Docker or GKE cluster deployment, monitor it via Prometheus/Grafana, and summarize latency/throughput/delivery results. Use when the user says 'run the network benchmark', 'network stress test', 'apollo_network_benchmark', 'p2p benchmark', or wants to benchmark gossipsub/sqmr broadcasting.
- ▌ Investigate Echonet Resync · starkware-libsInvestigate why an Echonet resync triggered. Use when someone mentions an Echonet resync, a failing block in an `echonet-*` namespace, or a `block_hash_mismatch` / `transaction_commitment_mismatch` / `echonet_only_revert` / `gateway_error`. The skill sets up port-forwards, prefers GCP Logs Explorer for history, cross-checks the produced block against mainnet, and recognizes recurring failure patterns before reporting back.
- ▌ Fri Protocol · starkware-libsCircle FRI protocol specifics for STWO: commitment phase, query phase, folding operations, security parameter derivation, and multi-step folding. Use when modifying FRI prover or verifier code, changing FRI parameters, or reviewing folding operations.
- ▌ Debugging Zkp · starkware-libsDebugging failing proofs and constraint violations in STWO. Covers failure mode diagnosis by symptom, FRI error triage, channel desync detection, and distilled theory cross-references. Use when prove() fails, verify() rejects a valid proof, constraints don't hold on a trace, or logup sums don't balance.
- ▌ Testing Strategy · starkware-libsTest taxonomy, coverage strategy, and test patterns for STWO. Covers unit test locations, coverage gaps, property testing opportunities, CI matrix, and patterns for field, prove-verify, and constraint tests. Use when adding tests, reviewing coverage, debugging failures, or assessing test adequacy.
- ▌ Zk Stark Foundations · starkware-libsSTWO-specific STARK architecture and protocol flow. Provides implementation locations, security parameter configuration, proof flow mapping, and invariants. Use when working on proof system code, reviewing constraint logic, modifying FRI parameters, or auditing soundness.
- ▌ Finite Field Arithmetic · starkware-libsM31/CM31/QM31 field arithmetic for STWO. Covers reduction logic, SIMD field implementations, extension field embedding, and batch inverse. Use when modifying code in core/fields/, prover/backend/simd/m31.rs, cm31.rs, qm31.rs, or reviewing any field operation changes.
- ▌ Circle Stark Mathematics · starkware-libsCircle-group-specific mathematics for STWO: circle points, cosets, domains, circle FFT, circle polynomials, twin cosets, vanishing polynomials, and the M31 circle group structure. Use when modifying any code involving these concepts, which form the non-standard foundation distinguishing STWO from traditional multiplicative-subgroup STARKs.
- ▌ Performance Optimization · starkware-libsSTWO-specific performance patterns: SIMD backends, FFT hot paths, memory pooling, benchmarking, parallel proving, and regression detection. Use when benchmarking, optimizing hot paths, working on SIMD backends, analyzing regressions, or reviewing parallel proving code.
- ▌ Rust Codebase Conventions · starkware-libsRepo-specific Rust patterns for STWO: feature flags, error handling, type patterns, unsafe policy, no_std compatibility, formatting, and clippy configuration. Use when contributing code, reviewing PRs, or understanding code style decisions.
- ▌ Security Review Checklist · starkware-libsSecurity review checklist for STWO. Separate from soundness review. Covers: side-channel surface, input validation, unsafe code audit, dependency security, API misuse patterns, and proof malleability. Run this for any change affecting the public API, proof format, hash functions, or memory-safety-critical code.
- ▌ Air Constraint Engineering · starkware-libsDefines how to audit and modify AIR constraints in STWO. Covers Flat AIR model, constraint degree bounds, EvalAtRow trait, LogUp interactions, constraint debugging, and soundness criteria. Use when working on constraint definitions, the constraint framework, logup interactions, or any component's constraints.
- ▌ Soundness Review Checklist · starkware-libsStructured checklist for reviewing soundness-critical code changes in STWO. Run this checklist before approving ANY modification to: constraint logic, FRI protocol, verifier, field arithmetic, polynomial commitment scheme, Fiat-Shamir channel, proof serialization, or security parameters.