architecture — independently scanned and version-tracked by SaferSkills.
SaferSkills independently audited architecture (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.
Rust for memory safety without GC. Stable toolchain for reproducible builds.
Faster verification (~3x), no signature malleability, simpler implementation. Trade-off: less Bitcoin ecosystem tooling compatibility.
BLAKE3 for speed in Merkle trees (internal). SHA-256 for block headers to maintain RandomX compatibility for Phase 2 mining.
No floats anywhere in consensus-critical code. Fixed-point u64 with 10^8 precision. Prevents platform-dependent rounding differences.
Gossipsub for block/tx propagation. Kademlia for peer discovery. Noise for encrypted transport. Mature Rust implementation.
Column families: blocks, headers, utxos, peers, meta. Provides atomic writes across families and efficient prefix scans.
Compact binary serialization. Deterministic encoding for consensus. ~10x faster than JSON, ~3x smaller.
Phase 1 uses simple hash-prefix PoW for fast iteration. Phase 2 switches to RandomX for ASIC resistance.
Genesis block allocates 5% of max supply to development fund. Vests linearly over 4 years.
rill1<base58check(sha256(ripemd160(pubkey)))> Human-readable prefix for identification. Double-hash for collision resistance. Base58Check for error detection.
Status: Accepted (2026-02-22) Spec: PROOF_OF_CONDUCT_SPEC.md | Review: PROOF_OF_CONDUCT_REVIEW.md
AI agent wallets need on-chain reputation that has economic teeth. Existing solutions (ERC-8004, Coinbase Agentic Wallets) decouple reputation from economics. RillCoin's L1 decay primitive uniquely enables tying conduct directly to wealth erosion.
Adopt Proof of Conduct as a native L1 extension. The concentration decay rate becomes dynamic per agent wallet, controlled by a Conduct Score (0–1000) that derives a Conduct Multiplier applied to the base decay rate.
u64), not f32. Variance comparison replaces stddev (no sqrt). Smoothing uses (85 * old + 15 * raw) / 100.TxType enum on Transaction for agent operations (RegisterAgent, AgentContract*, PeerReview, Vouch/Unvouch, UndertowDispute). No backwards-compat concern pre-mainnet.effective_decay_rate capped at DECAY_PRECISION to prevent overflow under Undertow (10× multiplier).agent_wallets per ADR-006.rill-core (foundation types, traits, errors, constants)
↓
rill-decay (concentration decay algorithm, cluster index)
↓
rill-consensus (block validation, chain selection, UTXO management)
↓
rill-network (P2P, libp2p, message relay)
↓
rill-wallet (keys, addresses, transaction construction)
↓
rill-node (full node binary, RPC, storage, main loop)~30 seconds. Free. No account. Every finding cites a rule and a line of evidence.