Rust low-latency Solana arbitrage execution skill for the Solana AI Kit: Yellowstone gRPC ingestion, simulate-first Jito bundles, tx landing, risk/kill-switch. MIT.
SaferSkills independently audited solana-arb-executor (Agent Skill) and scored it 100/100 (green). The audit ran 55 deterministic rules across Security, Supply Chain, Maintenance, Transparency, and Community; it found 0 high-severity and 0 lower-severity findings. The full rule-by-rule trace and per-finding evidence are below. Free, methodology-open.
Findings & checks · 0 flagged
Every scanned point with the score it earned and what moved between them.
First recorded scan — no prior version to compare against.
The primary manifest — the file an agent reads to learn what this artifact does.
Scaffolds the Rust execution hot path for Solana arbitrage: stream -> detect -> simulate -> bundle -> land, wrapped in a risk gate (caps, circuit breaker, kill-switch) and a confirm-gate. This is the on-chain layer that prior art only analyzes: agiprolabs/claude-trading-skills is Python analysis (signals, backtests, no execution), and TypeScript arb skills punt the latency-sensitive landing path. This is the genuine white space -- the part that has to be fast, correct, and safe.
Honest economics (load-bearing): spatial DEX arb on Solana is infrastructure-dominated and heavily contested; raw price gaps go to parties with better network position than a single box. A retail operator's durable edge is not raw speed -- it is correct landing (simulate-first, right CU price, tip sizing, retry discipline), risk discipline (caps/breaker/kill-switch), and less-contested opportunities (funding-basis, cross-venue rate gaps, longer-horizon plays). This is scaffolding + knowledge, not a hosted bot and not a guaranteed profit machine.
DRY_RUN=true, REQUIRE_CONFIRM=true. Always simulate before submit.references/sdk-versions.md (verified 2026-06). Flag jito-sdk-rust as early 0.x.../solana-dev/ via references/delegation.md. Do not duplicate it here.Map the user's intent to exactly one leaf.
| User asks about... | Open this leaf |
|---|---|
| hot-path design / latency budget | references/architecture.md |
| stream accounts/slots / ShredStream ingestion | references/streaming-ingestion.md |
| Jito bundle / tip / leader timing | references/jito-bundles.md |
| land tx / priority fee / simulate / retries | references/transaction-landing.md |
| detect opportunity / spread / threshold | references/opportunity-detection.md |
| risk / circuit breaker / kill-switch / caps | references/risk-and-killswitch.md |
| safety / confirm-gate | references/safety-rails.md |
| crate / version / install | references/sdk-versions.md |
| program / Anchor / signing | references/delegation.md |
skill/templates/.templates/src/fees.rs and templates/src/risk.rs are std-only and self-contained -- run rustc --edition 2021 --test fees.rs (and risk.rs) standalone. The full crate (stream.rs, jito.rs, detector.rs, main.rs) compiles with cargo build against the pinned crates.../solana-dev/ for program / tx / signing.| Agent | Role | Model |
|---|---|---|
agents/arb-execution-architect.md | Designs the latency budget, ingestion/landing topology, and risk envelope before code | opus |
agents/rust-perf-engineer.md | Wires stream -> detector -> RiskGate -> simulate-first Jito bundle | sonnet |
| Command | Does |
|---|---|
commands/scaffold-executor.md | Emit the executor crate skeleton (DRY_RUN + REQUIRE_CONFIRM defaults) |
commands/wire-yellowstone.md | Generate the Yellowstone gRPC ingestion module with reconnect/backoff |
commands/add-killswitch.md | Drop in the RiskGate with caps + circuit breaker + kill-switch |
~30 seconds. Free. No account. Every finding cites a rule and a line of evidence.