A practical configuration kit for Claude Code, Codex, and other coding agents. It contains architectural principles, enforcement hooks, skills, drop-in rules, starter templates, and dynamic-workflow commands. Drop the relevant parts into a project so the agent starts from verified working patterns instead of rediscovering them every session.
This is not a collection of tips. It is a system that teaches your agent how to work - when to use one agent vs many, how to verify its own output, how to manage context across long sessions, how to not get poisoned by malicious packages.
Write-ups of the incidents that produced a rule or a hook here. Each states what was measured, what is inference, and what the fix does not cover.
| Starting from what you remember | Why a brand-new project arrives years out of date, why an invented package name is now a security problem rather than a 404, and where a dependency check has to sit to catch either. Ships as hooks/dependency-currency-guard.py. |
| Why an agent circles instead of acting | Two agents, identical rules, different gate shapes. Describing a fix instead of applying it turned out to be the only move with no gate on it. |
| Gates that cannot bootstrap themselves | A check that only arms once the thing it checks for already exists will never arm. The failure looks exactly like compliance. |
| Nine skills, one skeleton, and nobody reaching for them | Nine architecture skills existed and one was reachable — the one arguing for less code. How a one-sided advisory becomes a ratchet toward monoliths, and why filing knowledge by source book makes it unreachable. |
Three paths depending on what you need:
claude plugin install https://github.com/AnastasiyaW/claude-code-configThen in your Claude Code chat:
Read AGENTS.md and pick the principles, hooks, and skills that match my project.
git clone https://github.com/AnastasiyaW/claude-code-config ~/claude-code-config
# Copy the always-on safety hooks to your global config
python ~/claude-code-config/scripts/install_hooks.py --global
# Claude Code: copy a selected skill directory, not its parent category
mkdir -p ~/.claude/skills
cp -r ~/claude-code-config/skills/ai-ml/ml-research-lab ~/.claude/skills/
# Codex desktop: sync all public skills with backups for changed local copies
python ~/claude-code-config/scripts/sync_skills_to_codex.py --apply~/.claude/hooks/ stores the hook scripts; ~/.claude/settings.json is where they are registered. The install script merges safe defaults into your existing settings.
cd /your/project
git clone https://github.com/AnastasiyaW/claude-code-config .claude-config
python .claude-config/scripts/install_hooks.py --local
cp -r .claude-config/skills .claude/skillsThis keeps everything under .claude/ in your repo, nothing global.
| Project type | Minimum viable set |
|---|---|
| Any project | 5 safety hooks (destructive-command, secret-leak, git-destructive, git-auto-backup, session-drift-validator) + Principles 09 (Supply Chain), 10 (Agent Security), 11 (Documentation Integrity) |
| Web app | above + frontend-design skill + Principles 04 (Deterministic Orchestration), 05 (Structured Reasoning) |
| ML / data pipeline | above + flux2-*, diffusion-engineering, vlm-segmentation skills + Principles 03 (Autoresearch), 12 (Low-Signal Training) |
| Multi-agent / parallel sessions | above + mclaude + Principles 01 (Harness), 06 (Multi-Agent), 18 (Multi-Session Coordination), 19 (Inter-Agent Communication) |
| Library / package | above + Principles 08 (Skills Best Practices), 17 (DBS Skill Creation) |
| More than one CLI agent (Claude + Gemini / Codex) | above + rules/cross-harness-agents-md.md (one AGENTS.md per project, no symlinks) + gemini-delegate skill |
See AGENTS.md for the procedure an agent follows after install, HOW-IT-WORKS.md for the mechanics of each layer, and docs/runtime-wiring.md for the live verification contract.
Moved the config to a new machine or account and most skills stopped being
offered? Nothing raises an error when that happens — see
docs/skill-tree-recovery.md and run
python scripts/recover_skill_trees.py --report.
Architectural Principles - each one prevents a specific failure mode observed in real agent workflows:
- Self-evaluation bias? Separate Generator and Evaluator agents (Harness Design)
- Agent claims "done" but it's broken? Require durable proof artifacts (Proof Loop)
- Need to improve a prompt/skill/config? Automated Read-Change-Test loop (Autoresearch)
- LLM skips steps in complex workflows? Shell scripts for mechanical tasks, one step at a time (Deterministic Orchestration)
- Wrong debugging conclusions? Structured Premises-Trace-Conclusions format (Structured Reasoning)
- Task too big for one agent? Coordinator + specialized sub-agents (Multi-Agent Decomposition)
- Context degrades in long sessions? Treat CLAUDE.md as runtime config, not docs (Codified Context)
- Supply chain attack? Two config lines block packages younger than 7 days (Supply Chain Defense)
- Prompt injection via repo/MCP/web? Six-layer defense with real CVEs (Agent Security)
- Docs reference files that no longer exist? SessionStart hook validates every reference (Documentation Integrity) - ships with a working validator script
- Multi-agent infrastructure overhead? Separate brain from hands with lazy provisioning (Managed Agents)
- Agent cuts corners on critical rules? Absolute prohibitions with incident history (Red Lines)
- Long-running project lost its history? Condensed timeline per project, alongside handoffs (Project Chronicles)
- Skill is a monolithic wall of text? Split into Direction, Blueprints, Solutions (DBS Framework)
- Parallel chats fight over GPUs or overwrite each other's state? Append-only handoffs + lock-file coordination (Multi-Session Coordination)
- One chat needs to send a specific request to another? File-based mailbox with email-style threading and delivery receipts (Inter-Agent Communication)
- AI-assisted code review findings get rediscovered next PR? Review finding → regression test → invariant → cross-reference (Knowledge Base Enforcement)
- Zero-day vulnerabilities buried in source tree? LLM + rules + SAST pipeline (Vulnerability Detection Pipeline)
- User needs to choose between visual options (UI, design, diagrams)? HTML fragment server + file-based event queue (Visual Context Pattern)
- *Output keeps reverting to generic defaults (Inter font, SELECT , etc.)? Anti-attractor procedure + three-layer enforcement (Anti-pattern as Config)
- Merge conflict resolved "by logic" and lost half the work? Two-agent isolated reconciliation + verified-data priority (Merge Conflict Resolution)
- Built a coordination primitive from scratch? Map it to the classical analog first (Chubby lease, WAL, SMTP) and inherit 30 years of failure-mode literature (Coordination Primitives Mapping)
- Bug fix detoured into "this was already broken before me"? Five valid deferral reasons + mandatory durable proof artifacts (No-Pre-Existing Evasion)
- Long-run project's scope and progress scattered across 30+ handoffs? Three-artifact harness (PROBLEMS.md + feature_list.json + init.sh) with WIP=1 invariant and L1/L2/L3 evidence requirements (Feature Tracking)
- Feature rationale evaporates into git log after 6 weeks? Three-tier KB (Global -> Layer -> Feature narrative) with ULTRAPACK-style task.md, auto-allocated F-NNN ID, hyperlinked invariants (Feature-Layer Architecture)
- Model collapses to "predict zero" on residual/delta tasks? Traps and fixes for low-signal training (overlay maps, denoise deltas, color-correction residuals), from 4 rounds of real failure (Low-Signal Residual Training)
- Deep research results evaporate with the conversation? Save structured findings to an incoming folder -> review -> knowledge base pipeline (Research Pipeline)
- Need a human-browsable memory view without a second source of truth? Use an optional Obsidian-compatible Markdown hub over the private archive (Obsidian Mind adoption note)
- Building a brand-new agent and not sure what to decide first? 15-section MVP blueprint: autonomy level -> tool risk classes -> permission matrix -> budgets -> evals -> release checklist (MVP Agent Blueprint)
Need smaller diagnostic command output? The optional RTK integration is
pinned, checksum-verified, fail-open, and tested separately from safety hooks.
See docs/rtk-integration.md and
scripts/rtk_integration.py; it is never a substitute for raw evidence.
Ready-to-use hooks that enforce rules mechanically, not probabilistically (install via scripts/install_hooks.py; full map with bypass keys in rules/safety-hooks.md):
| Hook | Event | What It Does |
|---|---|---|
| session-drift-validator | SessionStart |
Validates file references in CLAUDE.md at session start |
| destructive-command-guard | PreToolUse |
Blocks rm -rf, git push --force, DROP TABLE |
| secret-leak-guard | PreToolUse |
Prevents committing API keys, tokens, passwords |
| session-handoff-reminder | Stop |
Reminds to write handoff before closing long sessions |
| session-handoff-check | SessionStart |
Shows recent handoffs from previous sessions (latest per project) |
| handoff-closure-audit-guard | PreToolUse |
Blocks handoff writes that lack a closure audit for the primary task and related/scope-adjacent tasks |
| stop-phrase-guard | Stop |
Detects behavioral-regression phrases (ownership dodging, permission-seeking, premature stopping, deferral-via-"what next?") |
| keyword-skill-router | UserPromptSubmit |
Detects natural-language keywords and suggests matching skills (bilingual RU/EN) |
| api-key-leak-detector | PostToolUse |
Scans tool output for exposed API keys, tokens, secrets |
| command-injection-guard | PreToolUse |
Blocks shell substitution with non-trivial commands |
| git-destructive-guard | PreToolUse |
Blocks git reset --hard, push --force, branch -D |
| git-auto-backup | PreToolUse |
Creates backup branch before destructive git operations |
| self-harm-guard | PreToolUse |
Prevents agent from killing its own process, locking SSH, bare reboot |
| test-muting-guard | PreToolUse |
Blocks adding @skip, .only(), @Ignore to existing tests |
| backup-retention-cleanup | Stop |
Cleans up old backup branches (14-day retention) |
| file-cohesion-guard | PreToolUse |
Advisory: warns when a durable file is written to a scratch location (home root, Desktop, Downloads, /tmp) instead of the project structure |
| human-confirmation-guard | PreToolUse |
Requires explicit user confirmation before any deletion-intent command |
| ask-question-guard | PreToolUse |
Blocks deferral/menu AskUserQuestion ("what next?", "which of these?") on reversible work — decide and proceed instead |
| over-engineering-advisor | PostToolUse |
Advisory nudge when an edit adds a large code block or a new dependency — "is this the minimal solution?" (never blocks) |
| module-shape-advisor | PostToolUse |
The mirror of the row above: advisory nudge when the whole FILE has outgrown its shape — "where is the seam?" Fires on cumulative size, not on your edit, because that is how a file gets there (never blocks) |
| dependency-currency-guard | PreToolUse |
Blocks a manifest edit that names a package which does not exist, is too new or too little used to be a real recall (the slopsquat profile), or pins a fast-moving library far behind current |
| pre-push-public-repo-scan | git pre-push |
Two independent scans — regex and semantic — of a push to a PUBLIC repo; either one alarming blocks it. Private repos skip. Host and script names load from a local list, never from this file |
| shape_common | (library) | Not a hook: the one definition of "what shape is this file in", shared by module-shape-advisor and scripts/architecture_audit.py so the two cannot answer differently |
| harness-load-advisor | Stop |
Notices when a closing message reports a high-cost gate (production signing, a release VM) blocking work whose actual risk is lower, and says so. A feedback guard, not a bypass — it never lifts the gate |
| repeated-attempt-guard | PreToolUse + PostToolUse |
Stops the guess-and-retry loop: advisory on the third failed attempt at the same target, blocking on the fourth, unless something has been read since the last failure. One Read clears it — the block is lifted by the action that would have solved it three attempts earlier. Needs both events: PostToolUse records outcomes, PreToolUse decides |
| activity-journal-guard | PreToolUse |
Enforces the shared activity journal — blocks a mutating command on a tracked shared resource that does not log to its journal |
| coord-claim-guard | PreToolUse |
Claim-before-edit gate for multi-session / coord-enabled repos (blocks editing a file without an active claim) |
| continuity-contract-guard | PreToolUse |
Protects Claude/Codex continuation: no silent whole-file Write, out-of-scope edits, or near-whole-file replacement |
| continuity-session-check | SessionStart |
Surfaces the shared .claude/continuity/CONTINUITY.json contract and its preserve/do-not-redo decisions |
| cyrillic-bash-guard | PreToolUse |
Blocks raw non-ASCII (Cyrillic/CJK) in Windows Bash commands — encoding-corruption guard |
| feature-list-validator | Stop |
Validates feature_list.json discipline (WIP=1; done needs evidence) — companion to problems-md-validator |
| handoff-resume-gate | SessionStart |
Resume freshness-gate — complements session-handoff-check by gating on stale/unacknowledged handoffs |
| long-run-detector | SessionStart |
Auto-detects a long-running project and nudges adopting the [LONG-RUN] harness (feature_list.json / init.sh) |
| verify-deleted-guard | PostToolUse |
Verifies a destructive operation actually completed (object really gone) |
| db-snapshot-guard | PreToolUse |
Auto-snapshots the database before bypassed destructive SQL |
| claude-attribution-guard | PreToolUse |
Blocks commits/PRs carrying Co-Authored-By: Claude footers (see rules/no-claude-attribution.md) |
| pre-push-claude-attribution | git pre-push |
Final attribution gate before commits reach the remote |
| precompact-handoff-guard | PreCompact |
Demands a fresh handoff before context compaction; writes an AUTO-DRAFT fallback if none exists |
| test-gate-stop-hook | Stop |
Selects fast/integration evidence by Git-visible risk and blocks closing while selected tests are red or unproven |
| problems-md-validator | Stop |
Blocks closing with OPEN problems lacking a valid deferral reason |
| task-inbox-show | SessionStart |
Surfaces pending tasks from .claude/task-inbox/ |
| plan-gate | UserPromptSubmit |
Non-blocking nudge: substantive build/refactor ask + no plan artifact in the project -> one-line "freeze acceptance criteria first" reminder (max once/day) |
Supporting hooks and shared utilities (wire these when the project needs the corresponding workflow):
| Hook | Event | What It Does |
|---|---|---|
| conversation-history-capture | Stop |
Archives the local session transcript for searchable continuation |
| directory-creation-guard | PreToolUse |
Applies lifecycle labels and placement checks to new directories |
| docs-staleness-guard | SessionStart |
Surfaces stale project guidance before work begins |
| feedback-pending-show | SessionStart |
Shows queued corrections waiting for review |
| git-source-gate | Stop |
Checks that durable work is represented in Git before closure |
| github-workflow-security | PreToolUse |
Adds a security checklist before editing GitHub Actions workflows |
| kb-validate-gate | Stop |
Runs the project knowledge-base validator when opted in |
| session-feedback-capture | Stop |
Queues durable correction notes without blocking session closure |
| safety_common.py | shared | Shared event parsing and decision helpers for opt-in hooks |
Starter templates for common project types: web-app, ML project, library, code review, project chronicle, memory files, memory reference, proof plan, bug-fix prompt (anti-"pre-existing" constraints baked in), long-run project harness pack (drop-in feature_list.schema.json + feature_list.template.json + init.sh.template for any project crossing 5+ features and 5+ sessions). |
Dynamic workflow commands (workflows/) - ready-to-drop .js orchestration scripts for Claude Code dynamic workflows (/deep-review-flow, /research-cn-ru) plus EFFECTIVE-AGENTS.md - measured cost lessons (one agent() ≈ 95-150k tokens; resume as the main economy lever).
Cross-harness setup (rules/cross-harness-agents-md.md) - share one AGENTS.md per project between Claude Code, Gemini CLI, and Codex without symlinks: Claude imports it via @AGENTS.md, Gemini reads it via context.fileName, Codex natively. Companion skill gemini-delegate covers multi-account Gemini CLI delegation (quota ladders, account switcher scripts/gemini-switch.sh, trust boundaries).
For serial Claude/Codex handoff, use the cross-harness-continuation contract. It records the Git baseline, claimed files, accepted decisions, rejected approaches, and verification. The guard blocks silent rewrites and scope drift; an intentional redesign must use an explicit, reasoned replan mode.
Your agent picks the approach that fits. The alternatives/ directory compares 2-5 approaches for each problem, with pros, cons, and "when to choose" guidance:
| Problem | Approaches Compared |
|---|---|
| Multi-step orchestration | Harness Design, Proof Loop, Deterministic Orchestration, Prompt-only |
| Code review | Sequential checklist, Parallel competency, Cross-model, LLM + static |
| Iterative optimization | Autoresearch, HyperAgent, Manual, Eval-driven |
| Codebase scoping before changes | Belief Map / Code Graph, Symbol Index / LSP, Targeted rg, Full Context Upfront |
| Context in long sessions | JIT Loading, Full Context Upfront, Compaction, Fresh Sessions |
| Session transitions | Manual HANDOFF.md, Auto hooks, Session Journal, ContextHarness, Memory |
| Reasoning-quality regression | Config reset, Stop-phrase guard, Metric monitoring, Fresh-session A/B, Proof Loop |
If you have a project that crosses 5+ features and 5+ sessions of work, three drop-in artifacts close the gap that PROBLEMS.md + handoffs + chronicles alone leave open:
| Artifact | Question it answers | Where |
|---|---|---|
init.sh |
Is the project healthy right now? (binary check, <3 min target) | templates/long-run-project/init.sh.template |
feature_list.json |
What features exist and what state are they in? (machine-readable) | templates/long-run-project/feature_list.schema.json + .template.json |
PROBLEMS.md |
What is broken right now? Recovery procedures? | Already covered in rules — pairs with the two above |
Hard rules attached to this pack:
- WIP=1: at most one feature in
status: "in-progress"at any time - L1+L2+L3 evidence:
status: "done"requiresevidencefield referencing Syntax/Static + Runtime + System artifacts (durable files, not "tests pass" claims) doneis one-way: regression becomes a new feature, never roll back- Durable source and docs: creating
feature_list.jsonopts the project into the Stop gates for a Git worktree withoriginplus an agent-facing KB that stays current. Scratch folders remain outside this boundary.
To audit whether your project needs this pack — and which subsystem to fix first — invoke the new harness-audit skill:
/harness-audit
or trigger phrases like "audit my harness", "score my CLAUDE.md", "is my project ready for long-run". The skill produces a 5-subsystem scorecard (1-5 per dimension), identifies the bottleneck, and outputs a prioritized 3-step improvement plan with effort estimates and pointers to the templates above. Read-only — no changes applied unless you approve.
See principle 27 - Feature Tracking for the full framework. Templates and concepts adapted from walkinglabs/learn-harness-engineering (MIT license), integrated with our existing Proof Loop, Multi-Agent Decomposition, and No-Pre-Existing Evasion principles.
For the agent (you): When this repo is connected to your project, you get access to all principles and skills automatically. Use them as decision frameworks - when facing a choice (one agent vs many? how to verify? how to manage context?), check the relevant principle or alternative comparison.
New: HOW-IT-WORKS.md - technical deep dive into how each technology actually works, with real measurements.
Structure:
principles/- standalone architectural principles. Read the one that matches your current problem.rules/- drop-in.claude/rules/files: always-on working discipline (no-guessing, finish-the-task, deletion-confirm, autonomy-risk-tiers, quality-code) plus a consolidated safety-hooks reference. Agent-harness design rules (tool risk taxonomy, budgets, evals, observability, trust labels) now live on-demand in theagent-harness-designskill.alternatives/- side-by-side comparisons of 2-5 approaches per problem. Pick the approach that fits.hooks/- ready-to-use Python hook scripts for safety guards, session management, and discipline enforcement. Wire them withscripts/install_hooks.py.workflows/- drop-in dynamic-workflow commands (/deep-review-flow,/research-cn-ru) + measured cost lessons.templates/- starter CLAUDE.md and REVIEW.md files for different project types, plus the kb-skeleton and long-run-project scaffolding packs.skills/- domain skills (AI/ML, frontend, iOS, code review, video, writing, operational tooling). Loaded on demand; the generated list is in skills/README.md.scripts/- utilities: hook installer, config validator, cross-reference checker, KV-cache stats, skills-lock generator, public-repo sync with privacy scanner, Gemini account switcher.skills-lock.json- reproducible lockfile with content hashes of every skill (regenerate viascripts/generate_skills_lock.py).CLAUDE.md- compact summary of all principles for global config.
Start with L1 for any project. Add L2 when tasks repeat and optimization matters. L3 only when solo agent is not enough.
| Level | Focus | Principles |
|---|---|---|
| L1: Foundational | Single agent, planning, tool use | Deterministic Orchestration, Structured Reasoning, Skills Best Practices, DBS Skill Creation |
| L2: Self-Evolving | Feedback loops, memory, optimization | Autoresearch, Codified Context, Proof Loop |
| L3: Collective | Multi-agent coordination | Harness Design, Multi-Agent Decomposition, Managed Agents, MVP Agent Blueprint |
| Cross-cutting | Security + Integrity | Supply Chain Defense, Agent Security, Documentation Integrity, Red Lines |
| Cross-cutting | Session + Project Continuity | Codified Context, Project Chronicles, Research Pipeline |
Based on three-level agentic reasoning taxonomy (arxiv 2601.12538, 2504.19678).
Two principles specifically address agent security:
Supply Chain Defense - most poisoned npm/PyPI packages are caught within 1-3 days. Two config lines create a 7-day buffer:
# ~/.npmrc
min-release-age=7# ~/.config/uv/uv.toml
exclude-newer = "7 days"Agent Security - covers 7 real attack categories with documented CVEs: in-code prompt injection, repo metadata poisoning, package metadata, MCP tool poisoning, web content injection, memory poisoning, sandbox escape. Includes a six-layer defense architecture.
When a Claude Code session gets long, or you want to continue tomorrow on a different machine, or your current chat predates any automation you've set up - just tell the agent to prepare a handoff.
Type one of these phrases and hit Enter:
prepare handoffsave context for new chatwrite handoffhandoff this session
The agent writes a handoff file with:
- What was the goal
- What got done
- What did NOT work (the most valuable part - prevents repeating dead ends)
- Current state (working / broken / blocked)
- Key decisions and why
- The single next step
Then it stops. Close the chat. Open a new one in the same directory. The new session reads the handoff automatically (if you set up the SessionStart hook) or you can paste the file as your first message.
Two storage modes - pick one:
| Mode | When to use | Storage |
|---|---|---|
| Single-file (default, simpler) | One chat at a time | .claude/HANDOFF.md |
| Multi-session (opt-in) | You run multiple Claude Code chats simultaneously on the same project | .claude/handoffs/<unique>.md + append-only INDEX.md |
Single-file works for ~80% of users. Switch to multi-session only if you've actually hit last-writer-wins data loss from parallel chats. See rule file for both protocols and principle 18 for the theory behind the multi-session append-only invariant.
Why a phrase and not a button: the trigger lives in .claude/rules/session-handoff.md as plain markdown. No plugin install, no settings file, no hook. Works in any Claude Code session immediately. This is essential for migrating existing sessions that were started before you configured anything.
Copy the ready-made rule file from rules/session-handoff.md into your project's .claude/rules/ (or ~/.claude/rules/ for global) and you're done.
For automation nerds: pair this with a Stop hook that blocks long-session closure until a handoff is written. See alternatives/session-handoff.md for all 5 approaches compared.
If you run parallel chats and they need to talk to each other (not just leave state), see principle 19 - Inter-Agent Communication. Mini decision tree:
Broadcast "I'm done, anyone continue" → handoff (principle 18)
Claim exclusive resource → lock file (principle 18)
Ask a specific other session to do X → mailbox/<name>/ (principle 19)
Announce a decision for all running chats → mailbox/all/ (principle 19)
Multi-turn reply chain → mailbox with in_reply_to threading
Skills are practical tools for specific domains. The complete list is generated
from live SKILL.md frontmatter, so it cannot silently fall behind the source:
skills/README.md. Verify it with:
python scripts/generate_skills_catalog.py --check
python scripts/generate_skills_lock.py --checkThese work well alongside the principles:
- gstack - dev workflow skills: /review, /qa, /ship, /investigate, /design-review
- hookify - git hooks generator for Claude Code
- Semgrep - static analysis, pairs with deep-review
- task-orchestrator - MCP task orchestration with dependency ordering
Principles are updated with new research findings, real-world incidents, and community patterns. Security sections track actual CVEs and attack chains. See UPDATES.md for the full changelog.
Freshness is mechanical, not aspirational: scripts/sync_public_config.py + sync-manifest.json run a manifest-driven one-way sync from the author's live ~/.claude into this repo - EOL-normalized diffing, an explicit deny-list for machine-specific files, and a privacy-marker scanner that blocks anything private from reaching the public tree (--scan-repo --strict runs before every push). If you maintain your own private-config/public-fork split, the same script works for you - edit the manifest.
- Fork the repo
- Add/improve a skill (
skills/<category>/<name>/SKILL.md) or principle (principles/) - Skill descriptions = triggers for the model, not human summaries. Include
## Gotchasfrom real failures - For principles or alternatives: open an issue first
面向 Claude Code 智能体的实战配置系统，包含架构原则、方案对比、技能、Hook 脚本、drop-in 规则和项目模板。
核心功能:
principles/- 独立架构原则，每个解决一个具体失败模式rules/- drop-in 规则（工作纪律、安全 Hook 配套文档；Agent 设计规则已移至agent-harness-design技能）alternatives/- 每个问题 2-5 种方案对比，附决策表hooks/- 即用型 Hook 脚本（安全防护、会话管理、技能路由），用scripts/install_hooks.py一键注册workflows/- 动态工作流命令（/deep-review-flow、/research-cn-ru）+ 实测成本经验templates/- 适用于不同项目类型的 CLAUDE.md 起始模板 + 验证计划、记忆、项目编年史和长期项目脚手架（feature_list.json + init.sh）skills/- 领域技能（AI/ML、视频制作、前端、iOS、写作、代码审查、验证、运维工具，包括harness-audit五子系统评估、workflow-orchestration和gemini-delegate跨 CLI 委派）- 跨 harness 支持：每个项目一个
AGENTS.md，同时供 Claude Code、Gemini CLI、Codex 读取（无需符号链接），见rules/cross-harness-agents-md.md
安装: claude plugin install https://github.com/AnastasiyaW/claude-code-config 或直接复制所需文件。
灵感来源: 部分设计理念受到中国工程社区的启发，包括红线(红线)模式、规范驱动开发(OpenSpec)、经验库模式。
Система конфигурации для Claude Code агентов: архитектурные принципы, сравнения подходов, навыки, hook-скрипты, drop-in правила и шаблоны проектов.
Что внутри:
principles/- принципы, каждый предотвращает конкретный тип отказаrules/- drop-in правила: рабочая дисциплина (no-guessing, finish-the-task, deletion-confirm, autonomy-risk-tiers, quality-code), консолидированный safety-hooks reference; правила проектирования агентов (risk taxonomy, budgets, evals, observability) теперь в скиллеagent-harness-designalternatives/- сравнение 2-5 подходов для каждой проблемы с таблицей решенийhooks/- готовые скрипты (safety guards, handoff, drift validator, keyword router, secret leak detection, backup retention, test/problems gates и др.), регистрация одной командойscripts/install_hooks.pyworkflows/- готовые dynamic-workflow команды (/deep-review-flow,/research-cn-ru) + замеры стоимости агентовtemplates/- стартовые CLAUDE.md + план верификации + шаблоны memory и хроник + long-run harness pack (drop-infeature_list.json+init.shдля проектов с 5+ фичами)skills/- доменные навыки (AI/ML, видео, фронтенд, iOS, письмо, код-ревью, верификация, операционные инструменты, включаяharness-audit,workflow-orchestrationиgemini-delegate— делегирование в Gemini CLI с мульти-аккаунтом)- Кросс-harness: один
AGENTS.mdна проект читают Claude Code, Gemini CLI и Codex (без симлинков) —rules/cross-harness-agents-md.md
Установка: claude plugin install https://github.com/AnastasiyaW/claude-code-config или копирование нужных файлов.
MIT