AI Agent & Automation Workflows

Canonical Claude Code authoring kit covering Skills, sub-agents, plugins, slash commands, hooks, memory, settings, sandboxing, headless mode, and advanced agent…

moai-adk-go best-practices reference for the 4 harness specialists (cli-template-specialist, quality-specialist, workflow-specialist, hook-ci-specialist).

moai-adk-go domain-patterns reference for the 4 harness specialists (cli-template-specialist, quality-specialist, workflow-specialist, hook-ci-specialist).

Local Self-Evolution Loop (LSEL) APPLY engine — the playback-only consumer of approved decision.json records that drives .moai/hooks/lsel-apply.sh for the GOOS-local PROPOSE→APPLY seam closure (SPEC-LSEL-LOCAL-EVOLUTION-001 M3). Reads an approved decision.json, validates the target against the frozen allowlist (.claude/lsel/frozen-allowlist.json), mechanically refuses execution-meta targets lacking a synchronous-approval marker, applies the referenced diff.patch, appends an apply-ledger.jsonl row, and commits one lsel- -tagged Conventional Commit on the feature branch. M3 scope: APPLY bypass closure only (the frozen Go applier stays frozen — its write-flag at internal/harness/applier.go:22 stays false; REQ-LSEL-003).

Local Self-Evolution Loop (LSEL) curator — the CLUSTER + drain engine for the GOOS-local PROPOSE→APPLY seam closure (SPEC-LSEL-LOCAL-EVOLUTION-001). Companion-offset drain of .moai/lessons-inbox.jsonl with a drain-side severity filter that drops the ~65% Bash-timeout/sandbox noise, eventkey clustering with a frequency gate, and a Generative-Agents-style 1-10 importance score. Candidates stage at .moai/state/lsel/clusters.json. M1 = drain only (NO PROPOSE, NO APPLY, NO memory/ writes).

moai-adk-go best-practices reference for the 4 harness specialists (cli-template-specialist, quality-specialist, workflow-specialist, hook-ci-specialist).

moai-adk-go local dev reference — version management/release process (sec 5), shell-script hook development (sec 7), build & dev commands (sec 10). Load only when performing these specific tasks.

moai-adk-go domain-patterns reference for the 4 harness specialists (cli-template-specialist, quality-specialist, workflow-specialist, hook-ci-specialist).

HARD i18n rules digest for the oss-docs harness specialists working on moai-adk-go README 4-locale set and the docs-site (adk.mo.ai.kr). Covers the canonical-locale chains, the 4-locale same-PR obligation, Mermaid TD-only, the no-emoji + icon-shortcode rule, emphasis-marker spacing, the URL blacklist, version SSOT, vercel.json redirect pattern, and the immutable Vercel binding. Loaded FIRST by every oss-docs specialist before any edit.

README 4-file synchronization procedure for the oss-docs harness: English README.md as primary source, ko/ja/zh derivation, the shared language-switcher header contract, section-order parity checklist, and the manual verification recipe (no linter exists for READMEs). Loaded by the content-author and locale-translator specialists for any README work.

docs-site structure map for the oss-docs harness structure-curator: exact paths and schemas for hugo.toml, per-locale content/ /meta.yaml, data/menu/main.yaml (4-locale name maps + icon values), the icon-to-SVG-case coupling in layouts/partials/menu.html, shortcodes, the FROZEN moai-brand.css, vercel.json redirect examples, and the known design-vs-guides divergence. Loaded by the structure-curator before any navigation or config edit.

Mandatory verify recipe for the oss-docs harness — the runnable exit gate every specialist executes before returning: warning-free hugo build, sitemap existence, URL-blacklist grep, Mermaid LR/RL direction grep, 4-locale file-existence and section-count parity, README 4-file heading parity, and body-emoji scan. All checks are inlined here because docs-i18n-check.sh and genmenu.py do not exist.

Claude Design handoff package specialist for /moai brain Phase 7. Assembles 5-file handoff bundle (prompt/context/references/acceptance/checklist) for paste-ready claude.com Design session. Handles brand-absent fallback and section regeneration.

Markdown-to-single-file-HTML report renderer. Six modes (status, incident, plan, explainer, financial, pr) selected by report type, crossed with three audience tiers (expert, basic, learn) derived from the active output style. The basic and learn tiers enrich the HTML with mermaid flowcharts, worked examples, and plain-language primers; the expert tier stays dense. Zero external JS/CSS framework dependencies — inline SVG charts, a font-CDN exception for Korean readability, and a tier-gated mermaid-CDN exception. Self-contained output for email attachment, print, and offline viewing.

AI text humanization and 윤문 (post-editing) specialist that detects and removes AI tells while preserving meaning, facts, and figures. Covers Korean, English, Japanese, and Chinese with a shared severity model (S1/S2/S3), quality grades (A/B/C/D), and 30%/50% over-editing guardrails. Use to make AI-generated text read as human-authored without changing what it says (de-ai, naturalness pass).

Ideation domain specialist: Lean Canvas assembly, SPEC decomposition list extraction, and Diverge-Converge pipeline for product proposal generation. Use during /moai brain Phase 2 (Diverge), Phase 4 (Converge), and Phase 6 (Proposal).

Market and ecosystem research specialist for /moai brain Phase 3. Executes parallel WebSearch + Context7 queries, handles tool failures gracefully, and produces structured research.md artifacts with cited sources and research limitations.

Author editable SVG technical infographics — architecture, flow, comparison, hierarchy — by computing the layout numerically before writing markup, then rendering a 2x PNG via headless Chromium. Carries a CJK-first font stack, a deterministic source lint, and mermaid-vs-SVG selection rules.

CLI/Template domain knowledge for moai-adk-go covering cobra commands, go:embed template system, YAML config, and template rendering pipeline.

Harness learning subsystem coordinator. Produces Tier 4 auto-update proposal payloads consumed by the orchestrator (which surfaces them via AskUserQuestion) and orchestrates Apply/Rollback flows. Triggers when harness learning proposals are pending or learning lifecycle management is needed.

MoAI-ADK harness pattern library — unified domain knowledge covering hook/CI dispatch (PostToolUse, SessionStart, GitHub Actions, release automation), workflow patterns (SPEC structure, EARS, MX tags, plan-run-sync pipeline), and Go quality gates (testing, linting, coverage, race detection, LSP). Use for moai-adk-go harness work — NOT for general MoAI agent patterns (see moai-foundation-cc).

Meta-harness skill that designs project-specific agent teams and generates the skills they use. Adapts the revfactory/harness 7-Phase workflow to MoAI's agent ecosystem. Triggered by /moai project Socratic interview and produces dynamic harness- skills + .claude/agents/harness/ + .moai/harness/ artifacts.

REST/GraphQL API design patterns, error handling conventions, and input validation reference for backend development. Agent-extending skill that amplifies expert-backend expertise with production-grade API patterns. Use when designing APIs, implementing endpoints, or reviewing backend code. NOT for: frontend development, DevOps, database schema design, security audits.

Cross-model audit convergence reference for the plan-auditor and sync-auditor agents. Documents how to invoke the auditmulti MCP tool to fan a code review out across the codex and GLM (z.ai) backends in parallel, converge their verdicts with the in-session Claude verdict, and fold the resulting per-backend verdicts + disagreement flag into the audit output. The single skill both audit entry points load — no duplication.