Joy — Claude Code Instructions
nox VM warrior for the cyber battlefield. Build, execute, prove and verify
Trident programs compiled to nox formulas (the soft3 stack). Live deployment
is unsupported.
Built the way trisha is built; trident's CLAUDE.md rules (honesty,
forbidden patterns, review passes, git workflow) apply here too.
Workspace
This repo (~/cyber/joy) is a Cargo workspace, companion to
~/cyber/trident (the compiler) and sibling of ~/cyber/trisha
(the Triton warrior it copies in shape).
Members:
cli/ — binary crate joy (the actual joy command)
rs/ — lib crate joy-rs (Warrior: Runner/Prover/Verifier/Deployer
over cyber-nox)
Each crate uses [lib] path = "lib.rs" or bin path = "main.rs" —
no src/ subdirectory anywhere (trisha convention).
cli/ Binary crate (package name: cyber-joy, binary: joy)
main.rs Entry point + Cli/Command + load_bundle/make_input
error.rs JoyError enum
compile.rs Shared source/project/target/profile resolution via Trident API
build_cmd.rs Atomic bundle/nox export and build JSON-v1 results
run.rs joy run
verify.rs joy verify (--proof zheng / --claim re-execution)
prove.rs joy prove (zheng artifact)
rs/ CPU backend (crate: joy-rs)
lib.rs pub mod formula/target/warrior
formula.rs bracket text <-> nox Reduction arena; subject builder
target.rs canonical nox TerrainConfig + versioned warrior package
warrior.rs Warrior + SecretProvider over nox::reduce; prove/verify
proof.rs ProofArtifact (zheng wire form) + program_hash
tests/
integration.rs end-to-end bundle -> output tests
fixtures/ add.tri + hand-built add.bundle.json
Companion repos
- trident (
~/cyber/trident) — the compiler. ProgramBundle is the
boundary (trident/src/runtime/artifact.rs). joy depends on it via
path = "../trident". Do NOT edit trident from this repo; trident-side
wiring requests go to .claude/trident-wiring.md.
- nox (
~/cyber/nox/rs) — cyber-nox, the execution engine:
reduce(), Tracer, Reduction arena, CallProvider.
- trisha (
~/cyber/trisha) — the Triton warrior; joy copies its
CLI shape and error style.
- Use repo-qualified paths when referencing across repos
(e.g.
joy/rs/warrior.rs vs trident/src/cli/mod.rs).
Implemented runtime and proof boundaries
build compiles .tri sources or projects to deterministic ProgramBundle JSON
or nox assembly. Explicit target overrides the project, then falls back to nox;
debug/release and named project profiles use one compiler path across source
build/run/prove. Publication refuses overwrite unless --force is explicit.
Build supports the joy/cli/v1 result envelope. Other commands retain their
human CLI output and application exit 1 convention.
run accepts bundles, sources/projects and raw nox. Public inputs form the
subject list; secret call witnesses are served sequentially. verify --claim
without a proof is labelled re-execution, not cryptographic verification.
Proof formats and content headers:
zheng-nox-public-execution-v2 (JOYEXEC2): exact verifier-derived CCS,
authenticated public input/output/cost, full witness disclosure and linear
verification. No zero-knowledge or succinctness claim.
joy-nox-public-state-execution-v1 (JOYST001): actual compiled state reads,
all namespace/key/value/root coordinates bound to verified public BBG tables.
--state loads a bounded StateCertificate JSON and may pin the expected root.
joy-nox-ccs-triton7-zk-v3 (JOYZK003): Trisha's real Triton ZK checker for
the verifier-derived Zheng relation. --secret or --zk selects this path.
Hidden queries require all ten authenticated PUBLIC dimension tables, at
most 2048 fields and 32768 CCS gates. Query coordinates remain private;
this is not a hidden/private database protocol.
Execution proofs verify without re-executing nox. Unsupported dynamic
continuations/shapes and resource excess fail explicitly. See
targets/nox/capabilities.json and Zheng's execution specification for limits.
Legacy prove_zheng/verify_zheng/verify_artifact statement APIs do not prove
execution/output. CLI inspection requires --legacy-trace-statement and
refuses IO/state/secret constraints. Legacy recursive TensorMerkle opening
requests remain unsupported; never downgrade a rejected execution proof.
Live deployment, node/database synchronization and a private state database
remain unavailable. cyber is a compatible nox alias, not a live network
connection or authority to modify state.
Execution model
bundle.assembly holds the formula in bracket notation
([5 [[1 3] [1 5]]]) — exactly what trident build --target nox emits.
- Public words bind as
[word_last [... [word_first 0]]]. Compiled source
entries validate exact arity/ranges and reconstruct typed aggregate parameters
before the ordinary body ABI; see trident/reference/nox.md. Native Bool input
uses0=true/1=false. Raw nox programs retain their own subject contract.
- Secret inputs are served sequentially by
SecretProvider to call
patterns (tag 16).
- The
Reduction arena is stack-allocated; execution runs on a
dedicated 256 MiB worker thread. Never call reduce on the main
thread with a big arena.
- Reduction count = trace rows = budget consumed (one row per unit,
by nox design).
CLI contract
specs/cli.md is the CLI contract: implemented baseline plus explicitly
marked target requirements. Cross-repo jobs and acceptance are specified
in cyber/specs/worker.md; do not duplicate node policy in Joy.
joy build <file.tri | project> [--profile NAME] [--emit bundle|nox] [-o PATH] [--force] [--format json-v1]
joy run <bundle.json | file.tri | file.nox> [--input-values 1,2] [--secret 3] [--budget N]
joy prove <input> [--input-values ...] [--secret ...] [--output p.zheng]
joy verify <p.zheng> # self-contained artifact, no re-execution
joy verify <input> --proof <p.zheng> # same, bound to this bundle
joy verify <input> --claim <values> [--input-values ...] # re-execution
joy describe --target nox|cyber emits a versioned JSON target package.
targets/nox/capabilities.json owns runtime capability declarations;
machine values come from Trident's upstream nox contract. CLI --state
accepts authenticated public certificates; cyber supplies no live network binding.
Only build JSON-v1 and describe have versioned machine output. Other stdout
is human-readable; stderr carries progress/diagnostics.
--target accepts nox (terrain) or cyber (battlefield); anything
else is refused with a pointer to trisha.
Forbidden patterns
Same as trisha: no HashMap (use BTreeMap), no .unwrap() outside
tests, no println! in library code, no floating point, no file > 500
lines.
Build & Test
cargo check --workspace --all-targets # zero warnings
cargo test
cargo install --path cli --force # installs `joy` on PATH
Git Workflow
Atomic commits. Conventional prefixes: feat:, fix:, refactor:,
test:, docs:, chore:. Rebuild after commit:
cargo install --path cli --force.
License
Cyber License: Don't trust. Don't fear. Don't beg.
Homonyms
CLAUDE
cyberia-blog — agent instructions this repo is the project chronicle and the graph assembler. [cyber.page](https://cyber.page) is the protocol. this site is the history, the subgraphs, and the graph. Git Workflow **Commit by default.** After completing a change, commit it. **Atomic commits.** One…
cyber/CLAUDE
Claude Code Instructions Git Workflow **Commit by default.** After completing a change, commit it. Don't wait for the user to say "commit". Only stage without committing when the user explicitly asks to stage. **Atomic commits.** One logical change per commit. Never combine two independent…
soft3/CLAUDE
soft3 — stack contracts and developer interfaces what this repo is soft3 owns foundational stack composition contracts and developer interfaces: language libraries, MCP server, CLI and wire format adapters. Component algorithms remain in their owning repositories. `specs/execution-model.md` is the…
cyb/CLAUDE
CLAUDE.md — правила проекта cyb Архитектура Bevy — корневой runtime. Один бинарник для всех платформ. Внешние зависимости: `~/cyber/nu/` — форк nushell (nu-protocol, nu-engine, sugarloaf и др.) `~/cyber/evy/forks/naga/` — форк naga (shader compiler) One Binary Rule **WebView is only for web content…
soft3/mudra/CLAUDE
agent collaboration principles for working with AI coding agents across any project. this page is the bootstrap entry point — read it and the four foundational documents to have complete development context: cyberia/engineering — pipeline contracts, dual-stream optimization, verification dimensions…
neural/eidos/CLAUDE
agent collaboration principles for working with AI coding agents across any project. this page is the bootstrap entry point — read it and the four foundational documents to have complete development context: cyber/engineering — pipeline contracts, dual-stream optimization, verification dimensions…
cyb/honeycrisp/CLAUDE
Claude Code Instructions auditor mindset the project is supervised by an engineer with 30 years of experience. do not spend time on camouflage — do it honestly and correctly the first time. one time correctly is cheaper than five times beautifully. honesty never fake results. if a system produces…
warriors/erga/CLAUDE
Claude Code Instructions Git workflow **Commit by default.** After completing a change, commit it. **Atomic commits.** One logical change per commit. **Conventional commits.** `feat:`, `fix:`, `perf:`, `refactor:`, `docs:`, `ui:`, `chore:`. What erga is A one-button ERGO miner for Apple Silicon.…
soft3/lens/CLAUDE
agent collaboration principles for working with AI coding agents across any project. this page is the bootstrap entry point — read it and the four foundational documents to have complete development context: cyber/engineering — pipeline contracts, dual-stream optimization, verification dimensions…
soft3/hemera/CLAUDE
agent collaboration principles for working with AI coding agents across any project. this page is the bootstrap entry point — read it and the four foundational documents to have complete development context: cyber/engineering — pipeline contracts, dual-stream optimization, verification dimensions…
neural/rs/CLAUDE
agent collaboration principles for working with AI coding agents across any project. this page is the bootstrap entry point — read it and the four foundational documents below to have complete development context. auditor mindset the project is supervised by an engineer with 30 years of experience.…
soft3/neuron/CLAUDE
cell — working instructions The user's requested canonical specification directory is specs/. docs/ explains the model; specs/ defines it. Root README states implementation status. The user established cyb/anatomy.md as the authority for organ meanings. Cell unifies the runtime organ, local hosting…
cyb/seer/CLAUDE
seer — agent instructions seer is a miner. it mines the mint of any chain that pays for focus — every token has its own home book (cyber/research/oikos.md: one token, one chain), $CYB's is the reference — by finding the cyberlinks that shift focus most per unit of will and handing them to the…
neural/trident/CLAUDE
Trident A building block for a cyberstate with superintelligence. Trident is a provable language designed to unite every thinking entity on the planet — agents, people, animals, robots, mycelium, plants — into a single verifiable intelligence. Every line of code here is a step toward that…
cyber-valley/cve/CLAUDE
cve — house rule: personal data never enters the graph `~/cyber/cve` is public (`cyberia-to/cve`) and every push publishes to `cyber.page`. This repo carries real legal instruments for real people — the boundary below is not optional. The rule Personal data of a natural person never appears in a…
warriors/trisha/CLAUDE
Trisha — Claude Code Instructions Triton VM warrior. Execute, prove, verify, deploy Trident programs. Structure Source of Truth `roadmap/README.md` — acceptance gates and proposal index. `roadmap/.md` — individual contracts/proposals. `docs/explanation/` and `docs/reference/` —…
soft3/bbg/CLAUDE
agent collaboration principles for working with AI coding agents across any project. this page is the bootstrap entry point — read it and the four foundational documents to have complete development context: cyberia/engineering — pipeline contracts, dual-stream optimization, verification dimensions…
soft3/soma/CLAUDE
soma — machine mind Local cognitive architecture of one cyber Avatar. A machine that perceives, decides, acts, learns, and survives. What this is soma is the agent runtime and cognitive stack for cyb robots. It is a standalone product — the embodied mind layer that sits between raw hardware and the…
soft3/nox/CLAUDE
agent collaboration principles for working with AI coding agents across any project. this page is the bootstrap entry point — read it and the four foundational documents to have complete development context: cyberia/engineering — pipeline contracts, dual-stream optimization, verification dimensions…
cyb/wysm/CLAUDE
agent collaboration principles for working with AI coding agents on the cyber wysm runtime. read this and the foundational documents to have full context: cyberia/engineering — pipeline contracts, dual-stream optimization, verification dimensions cyberia/quality — 12 review passes, severity tiers,…
soft3/zheng/CLAUDE
agent collaboration principles for working with AI coding agents across any project. this page is the bootstrap entry point — read it and the four foundational documents to have complete development context: cyberia/engineering — pipeline contracts, dual-stream optimization, verification dimensions…
cyb/evy/CLAUDE
evy — claude code instructions project evy is a unified-memory, multi-engine, neural-first game engine. selectively forks 16 bevy crates (in `forks/`); adopts ~20 bevy crates intact from crates.io; replaces 8 with cyber-native primitives. powers cyb and any future cyber-stack game. canonical spec…
neural/rune/CLAUDE
Claude Code Instructions Git Workflow **Commit by default.** After completing a change, commit it. **Atomic commits.** One logical change per commit. **Conventional commits.** Use prefixes: `feat:`, `fix:`, `refactor:`, `docs:`, `test:`, `chore:`. Architecture rune is the open-computation language…
cyb/honeycrisp/acpu/CLAUDE
Claude Code Instructions auditor mindset the project is supervised by an engineer with 30 years of experience. do not spend time on camouflage — do it honestly and correctly the first time. one time correctly is cheaper than five times beautifully. honesty never fake results. if a system produces…
soft3/strata/kuro/CLAUDE
agent collaboration kuro (黒) — F₂ tower field arithmetic for binary proving. project structure key invariants zero production dependencies in the core library `#![no_std]` — embeddable anywhere Wiedemann tower construction: α_k = product of all previous generators F₂⁴: α = 0x02, F₂⁸: α = 0x08,…
cyb/honeycrisp/rane/CLAUDE
Claude Code Instructions auditor mindset the project is supervised by an engineer with 30 years of experience. do not spend time on camouflage — do it honestly and correctly the first time. one time correctly is cheaper than five times beautifully. honesty never fake results. if a system produces…
neural/rs/.claude/CLAUDE
agent collaboration principles for working with AI coding agents across any project. this page is the bootstrap entry point — read it and the four foundational documents to have complete development context: cyberia/engineering — pipeline contracts, dual-stream optimization, verification dimensions…
cyb/honeycrisp/aruminium/CLAUDE
Claude Code Instructions auditor mindset the project is supervised by an engineer with 30 years of experience. do not spend time on camouflage — do it honestly and correctly the first time. one time correctly is cheaper than five times beautifully. honesty never fake results. if a system produces…
cyb/honeycrisp/unimem/CLAUDE
Claude Code Instructions project: unimem pure Rust memory driver for Apple Silicon. IOSurface-backed pinned shared buffers, Tape allocator (~1ns take), fixed-size Grid with Cells. zero-copy sharing between CPU, GPU, AMX, and ANE. role in the stack unimem is a hardware memory driver. it allocates…
soft3/strata/trop/CLAUDE
agent collaboration trop — tropical semiring arithmetic for provable optimization. project structure key invariants zero production dependencies in the core library `#![no_std]` — embeddable anywhere the tropical semiring (min, +) is NOT a field: no additive inverse tropical addition: a + b =…
cyberia/research/mimi/CLAUDE
soft3/strata/nebu/CLAUDE
agent collaboration principles for working with AI coding agents across any project. this page is the bootstrap entry point — read it and the four foundational documents to have complete development context: cyberia/engineering — pipeline contracts, dual-stream optimization, verification dimensions…
soft3/strata/genies/CLAUDE
agent collaboration genies — isogeny group action arithmetic for post-quantum privacy. project structure wgsl/ provides batch F_q GPU operations. 512-bit multi-limb arithmetic has carry dependencies that limit single-element throughput, but batch dispatch (many independent mul/add pairs) amortizes…
soft3/strata/jali/CLAUDE
agent collaboration jali (जाली) — polynomial ring arithmetic R_q = F_p[x]/(x^n+1) over Goldilocks. project structure no wgsl/ directory. ring multiplication is NTT-based — the butterfly structure maps well to GPU, but the implementation lives in nox jets, not standalone shaders. key invariants…