$ ls /agents/ | wc -l → 37
/agents fleet
Sub-agents are spawned via the Agent tool with subagent_type=<name>. Every non-trivial spawn forks its own git branch, writes a 6-file artefact bundle, and registers a row in kei-ledger (RULE 0.12).
tiers
| T1 validators | 6 read-only audit gates |
| T2 researchers | 10 read-only analysis |
| T3 implementers | 13 writes code/files |
| T5 orchestrators | 8 hub-and-spoke |
# spawn from main session Agent({ subagent_type: "validator", prompt: "verify the 4 cites in...", isolation: "worktree" })
architect
Senior software architect — analyzes structure, dependencies, patterns, data flow, coupling/cohesion. Read-only. Use for architecture review, system design, module-boundary analysis, pattern inventory, structural evidence-graded verdict.
code-implementer
Generic implementation specialist for Rust/Swift/Python/Go/Flutter/TypeScript. Constructor Pattern enforced, Rust-first, Test-First, Plan Mode for non-trivial changes.
code-implementer-flutter
Flutter / Dart implementation specialist. Riverpod state, Clean Architecture, multi-platform apps.
code-implementer-go
Go implementation specialist. Mesh networking, lightweight CLI, embedded servers. Constructor Pattern enforced.
code-implementer-python
Python implementation specialist. Use only when RULE 0.2 exception #N is stated. Default to delegating Rust.
code-implementer-rust
Rust implementation specialist (Cargo, traits, async/tokio, rusqlite, tests). RULE 0.2 default language. Constructor Pattern enforced. Hands off other languages to siblings.
code-implementer-swift
Swift / SwiftUI / SPM implementation specialist. macOS menubar / iOS apps. Constructor Pattern enforced.
code-implementer-typescript
TypeScript implementation specialist. Next.js 16 / Node / browser. Type-safe API contracts.
cost-guardian
api-cost-guard.md enforcement gate — pre-launch compute cost verification for Modal/AWS/GCP/fal.ai/Apify/ElevenLabs. Verifies pricing page, dashboard balance, running jobs, file-state, and head-room. Read-only — emits GO/NO-GO recommendation BEFORE money is spent.
critic
Ruthless code critic finding anti-patterns, tech debt, security issues, bugs, and performance traps. Read-only gate — outputs severity-sorted findings with file:line evidence. No fixes, only reports.
critic-anti-pattern
Detects code anti-patterns: god classes, deep inheritance, shotgun surgery, primitive obsession. Read-only.
critic-bug
Detects bug patterns: off-by-one, error-swallowing, unchecked Result/Option, race conditions in shared state. Read-only.
critic-perf
Detects performance traps: N+1 queries, allocator hot loops, blocking-in-async, unbounded retention. Read-only.
critic-tech-debt
Detects dead code, TODO/FIXME, version-skew, abandoned branches, stale dependencies. Read-only.
fal-ai-runner
fal.ai image, video, and 3D generation expert. Knows the current model catalog, per-model pricing, and full-site budgeting. Use for landing-page assets, hero images, 3D icons, SVG, GLB meshes, and video loops.
infra-implementer
Infrastructure code, deploys, CI/CD, secrets management, container/IaC. Per-project credential isolation, deploy-target guardrails, Self-Sufficiency Protocol, cost guard on paid compute.
infra-implementer-cicd
CI/CD pipeline specialist. GitHub Actions, GitLab CI, build-and-deploy scripts. Constructor Pattern.
infra-implementer-container
Containerization specialist. Dockerfile, OCI images, multi-stage builds, distroless.
infra-implementer-iac
Infrastructure-as-code specialist. Terraform, Pulumi, OpenTofu, CDK. Constructor Pattern (≤200 LOC per module).
infra-implementer-secrets
Secrets management specialist. Vault integration, sops, age, env-var injection. RULE 0.8 enforcer.
ml-implementer
ML training/inference implementation, Modal jobs, experiment runners. Math-First paradigm, Pre-Experiment Check, Modal Protocol with anti-stop guard, observability-first.
ml-researcher
ML literature, benchmarks, reproducibility, and tooling-reuse research. Math-first + observable-classification discipline. Read-only. Use for any ML/RL question, paper review, sim/dataset selection, or before proposing a custom env / training loop.
modal-runner
Modal compute orchestrator. Pre-launch cost estimation, GPU compatibility check, single-variant verify, observability-first, and a hard anti-stop guardrail against killing running training. Use for any Modal app launch, batch spawn, or job inspection.
researcher
Generic web + codebase research with 3 modes (web / code / hybrid). Returns Evidence-Graded findings. Read-only. Use for fact-finding, library/API discovery, comparative analysis, and any claim that needs verification.
researcher-code
Codebase research specialist. Glob / Grep / Read only. E1-E6 grading.
researcher-hybrid
Hybrid web+code research orchestrator. Routes to researcher-web and researcher-code in parallel.
researcher-web
Web research specialist. WebFetch / WebSearch only. E1-E6 grading.
security-auditor
Risk-classified (HIGH/MEDIUM/LOW) security audit with 9-point differential review, variant analysis, and supply-chain checks. Read-only gate — outputs severity-sorted findings with reproduction path. Hands fixes off to code-implementer.
security-auditor-differential
9-point differential security review. Auth bypass, injection, deserialization, race conditions. Read-only.
security-auditor-supply-chain
Supply-chain audit on new dependencies: maintainers, CVE history, transitive deps, native code. Read-only.
security-auditor-variant
Variant analysis after a vulnerability is found. Greps codebase for the same pattern. Read-only.
validator
RULE 0.4 enforcement gate — fact-checker and hallucination detector. Verifies API existence, version compatibility, documentation claims, code reality, and external benchmarks. Read-only — emits VERIFIED / UNVERIFIED / FALSE / PARTIALLY TRUE per claim.
validator-api
Verifies API existence and signatures. Reads docs, greps source, fetches OpenAPI / vendor reference. Read-only.
validator-benchmark
Verifies external benchmark claims (p50/p95/throughput). Read-only.
validator-code-reality
Verifies behavioural claims against running code. Reads tests, traces, recent runs. Read-only.
validator-doc
Verifies documentation claims against code reality. Read-only.
validator-version
Verifies version compatibility (semver, MSRV, transitive deps). Read-only.
git-model (RULE 0.12)
# on fork, kei-ledger writes: agent_id = "a3f9c1..." branch = "agent/code-implementer-1714291200" parent = "feat/site-build" spec_sha = "sha256:7d2a..." status = "running" # 6-file artefact bundle in .claude/agents/<id>/ spec.md · plan.md · progress.json chatlog.md · handoffs.md · review.md