The bug that breaks AI code in production

Agent writes a test.
Runs it on synthetic data it just generated.
Reports F1=1.000 ✅ SUCCESS.
You deploy.
Real-world data crashes everything.

This is called Validation Theater.
This is the most systematic Claude Code config for catching it automatically — enforcing Evidence Policy as deterministic Python hooks, not as instruction text.

Every claim carries an evidence marker —
[VERIFIED-REAL] (real data, sources cited) vs [VERIFIED-SYNTHETIC] (mock data, never valid for production claims).
Hard rule baked into rules/integrity.md: synthetic ≠ real.

_{Backed by 60 hooks · 114 skills · 15 agents + 3 teams · 1367 tests · 75% coverage · MIT · Deploy in 5 min}

📋 No install? Get the rules only:
docs/anti-hallucination.md — single file, ~500 tokens, paste into your CLAUDE.md.
Catches Validation Theater on its own. Compatible with any Claude Code config.

---

---

Why This Config?

Claude Code без этого конфига — как Ferrari на ручнике: мощный, но 60% потенциала потеряно. With this config — каждый коммит верифицирован, каждый агент помнит контекст, каждая ошибка записана и больше не повторяется.

Most configs are a single CLAUDE.md bloated to 3000+ tokens. This is different:

	Typical config	This config
Tokens/msg	3 000 – 5 000	~500 (core only)
Hallucinations	"trust me"	Evidence Policy + Confidence Scoring
MCP failures	session hangs	CircuitBreaker — auto-recovery in 60s
Prompt injection	no protection	InputGuard — 8 categories, auto-block
PII leakage	hope for the best	17 patterns: secrets + PII (email, phone, card, passport, СНИЛС)
Code review	optional	review-squad — parallel reviewer + sec-auditor
Permissions	ask for everything	PermissionRequest hook — 75% auto-approved
Agent memory	stateless	4 agents with persistent memory across sessions
Tests	"I'll write them later"	1367 tests, TDD-first, Test Protection hard rule

---

Where This Fits in the Ecosystem

Three repos solve three different layers. They compose, not compete:

pm-skills (phuryn)        → WHAT to build: discovery, strategy, PRD, assumptions
agent-skills (addyosmani) → HOW to build: spec → plan → build → test → review → ship
Claude-cod-top-2026 (you) → GUARDRAILS: evidence policy, anti-hallucination, hooks, security

	pm-skills	agent-skills	This config
Focus	Product management	Engineering lifecycle	Runtime guardrails
Primary user	PMs, founders	Engineers	Engineers on sensitive/research projects
Stars	~15k	~53k	early
Anti-hallucination	skill-level guidance	skill-level guidance	enforced in hooks — runs on every action
Evidence policy	none	none	[VERIFIED-REAL] vs [VERIFIED-SYNTHETIC] hard rule
Security hooks	none	none	PII redact, secrets guard, injection block
Multi-platform	Claude, Cursor, Gemini	Claude, Cursor, Gemini, Windsuff	Claude Code only

Use all three together for the full stack. Use only this one if you need runtime enforcement without PM tooling.

---

When to Use This vs everything-claude-code

everything-claude-code is a great alternative — bigger, multi-platform, Anthropic Hackathon Winner. Both are MIT, pick what fits.

Use everything-claude-code if:

• You want multi-language coverage (TS, Go, Java, Kotlin, Rust, C++, PHP — 12+ ecosystems)
• You work across multiple harnesses (Codex, Cursor, OpenCode, Gemini — not just Claude Code)
• You want a GUI dashboard for browsing components
• You like the paid tier path (ECC Tools GitHub App, free / pro / enterprise)

Use this config if:

• "Validation Theater" is a $$$ risk for you, not abstract — Evidence Policy is enforced as hard rule, not just a skill
• You work with sensitive data (PII, finance, healthcare) — built-in redaction hook scrubs sensitive strings before any external MCP call
• You need to read every hook before installing — only ~10 MB, plain Python, no JS dependencies, every file readable in 10 minutes
• You prefer Claude Code only with deep specialization over multi-platform breadth
• You speak Russian — README and rules have RU-first sections, useful for CIS dev teams

Comparison at a glance:

	everything-claude-code	This config
Surface	48 agents · 182 skills · 68 commands · ~31 MB	15 agents + 3 squads · 114 skills · 60 hooks · ~10 MB
Languages	TS, Py, Go, Java, Kotlin, Rust, C++, PHP, Perl	Python primarily
Harnesses	Claude Code, Codex, Cursor, OpenCode, Gemini, Antigravity	Claude Code only
Anti-hallucination	continuous-learning v2 with confidence scoring	Evidence Policy + Validation Theater Guard + Audit Verification Gate (synthetic ≠ real, enforced)
PII / sensitive data	generic	dedicated redaction hook + local-first (Ollama)
Audit Verification Gate	not in core	rules/audit-verification-gate.md — agent's [VERIFIED] = your [INFERRED]
Recurring mistake tracking	instinct-based	[×N] counter — after 3 occurrences a mistake becomes a hard rule
License	MIT (open core, paid GitHub App)	MIT (no paid tier)

If multi-language / cross-harness matters more than anti-hallucination focus — pick ECC. If anti-hallucination on sensitive data is your job-critical risk — pick this one.

---

What This Config Does NOT Do

Honest scope fence — to prevent misuse and save your time:

• ❌ Not multi-language. Python primarily. TS/JS supported via hooks but agents/skills are Python-tuned.
• ❌ Not multi-harness. Claude Code only. No Codex / Cursor / Gemini support — see everything-claude-code for cross-harness.
• ❌ Not a GUI / dashboard. Pure CLI + file-based config. scripts/hook_metrics.py is the closest thing (terminal table).
• ❌ Not a SaaS / managed service. Self-host only. No paid tier, no telemetry to author.
• ❌ Not for >50% of generic projects. This is optimized for anti-hallucination on sensitive data (PII, finance, healthcare, research). If your stack doesn't have validation theater risk, the overhead may exceed the value.
• ❌ Not a replacement for human review. Hooks catch deterministic mistakes (commits to main, leaked secrets, debug prints). Logic bugs still need Agent(reviewer) or human eyes.
• ❌ Not a methodology textbook. The rules (rules/*.md) document the patterns we use, but they're not a tutorial. Read docs/methodology.md for the explained version.

---

🚀 Start Here (pick your path)

New to this? Don't install everything at once. Pick the path that matches your goal:

Path	What you get	Time	Command
Evidence Only	[VERIFIED] markers + anti-hallucination	2 min	--profile=minimal
Daily Driver	+ 60 hooks + 15 agents + 114 skills	5 min	--profile=standard
Full Setup	+ MCP profiles + PII redaction + memory	10 min	--profile=full

Minimal path (recommended to start): installs just 3 files — CLAUDE.md, integrity.md, security.md. No hooks, no agents, no complexity. Add more when you need it.

---

Quick Start

# One-liner — Mac / Linux / WSL
git clone https://github.com/sergeeey/Claude-cod-top-2026.git && cd Claude-cod-top-2026 && bash install.sh --profile=standard --non-interactive

Windows (PowerShell): git clone https://github.com/sergeeey/Claude-cod-top-2026.git; cd Claude-cod-top-2026; bash install.sh --profile=standard --non-interactive

After install: restart Claude Code (/clear or new session) — hooks activate automatically.

Verify the stack

pytest tests/ -q --tb=short          # 1367 tests, 0 failures
bash tests/test_all.sh               # 296/296 smoke tests
ruff check hooks/ scripts/ tests/    # 0 errors

Plugin Install (recommended — Claude Code v2.1.80+)

# Register this repo as a marketplace source (once per machine)
/plugin marketplace add sergeeey/Claude-cod-top-2026

# Install the plugin
/plugin install claude-cod-top-2026

Windows note: Claude Code doesn't pre-register third-party marketplaces on Windows. Add to your ~/.claude/settings.json manually if /plugin marketplace add fails:

"extraKnownMarketplaces": {
  "claude-cod-top-2026": {
    "source": { "source": "github", "repo": "sergeeey/Claude-cod-top-2026" }
  }
}

Classic Install (all platforms)

git clone https://github.com/sergeeey/Claude-cod-top-2026.git
cd Claude-cod-top-2026

bash install.sh                                    # interactive
bash install.sh --link full                        # symlink + auto-update
bash install.sh --profile=full --non-interactive   # CI / headless

Profile	Installs	For whom
minimal	CLAUDE.md + integrity + security	Try Evidence Policy
standard	+ all rules + hooks + skills + agents	Daily work
full	+ MCP profiles + PII redaction + memory	Full control

---

60 Hooks — 25 Events

Hooks run 100% of the time — deterministic Python guards, not probabilistic instructions.

PreToolUse guards (12 hooks)

Hook	Protects Against
input_guard	Prompt injection via MCP (8 categories)
mcp_circuit_breaker	Session hang on MCP failure (auto-recovery 60s)
mcp_locality_guard	MCP call without local search first
pre_commit_guard	Commits to main · rm -rf · DROP TABLE
read_before_edit	Edit without prior Read
security_verify	Sensitive file edits without review
plan_mode_guard	3+ files edited without a plan
permission_policy	75% fewer permission prompts
checkpoint_guard	Risky ops without checkpoint

PostToolUse audit layer (11 hooks)

Hook	Protects Against
mcp_circuit_breaker_post	Records MCP failures for recovery
post_format	Unformatted code (ruff / prettier)
memory_guard	Forgotten memory update after commit
post_commit_memory	Context loss after commits
pattern_extractor	Lost lessons from fix: commits
drift_guard	Scope creep (NOT NOW keywords)
post_tool_failure	Repeated failures without strategy change
config_audit	Unauthorized settings changes
elicitation_guard	Elicitation events logging
moc_autolink	Notes written without Obsidian MOC links
observation_capture	Observations lost after file edits

Lifecycle · Session · Memory (20 hooks)

Hook	Event	Role
session_start	SessionStart	Load context from memory
session_end	SessionEnd	Trim + save state
session_save	Stop (async)	State persistence on exit
post_compact	PostCompact	Context reminder after compaction
pre_compact	PreCompact	Data preservation
keyword_router	UserPromptSubmit	Auto-trigger skills + power modes
thinking_level	UserPromptSubmit	Boost thinking for complex tasks
statusline	PostToolUse	Live bar: model · context% · cost
worktree_lifecycle	WorktreeCreate/Remove	Track experiment branches
agent_lifecycle	SubagentStart/Stop	Context in agent handoffs
subagent_verify	SubagentStop	Verify agent output quality
team_rebalance	TeammateIdle	Rebalance idle agents
stop_failure	StopFailure	Silent API error handling
task_audit	TaskCreated/Completed	Task event logging
instructions_audit	InstructionsLoaded	Track loaded rules
env_reload	FileChanged	Stale env after .env change
direnv_loader	CwdChanged	Wrong env after cd
async_wrapper	—	Non-blocking wrapper for bg hooks
webhook_notify	Stop (async)	Slack/Telegram on commit + session end
thematic_index_router	Stop	Route wiki entries to thematic indices

⚡ Power Modes — Magic Keywords

Type anywhere in your prompt:

Keyword	Mode	Effect
ralph	Persistent	Don't stop until done. Auto-retry. No confirmations.
autopilot	Full Autonomy	Plan + execute all steps. Only stop if truly blocked.
ultrawork / ulw	Max Parallelism	Launch agents concurrently. Batch ops. Speed > caution.
deep	Deep Analysis	Read everything. Evidence-mark all claims.
quick / быстро	Speed	Minimal output. No explanations. Just do it.

Modes are additive — ralph security audit = Persistent mode + security-audit skill.

---

13 Agents + 3 Teams

╔══════════════════════════════════════════════════════════╗
║  STRATEGIC — Opus          ·  20% of tasks               ║
║  navigator(memory:user)  architect  sec-auditor  teacher  ║
╠══════════════════════════════════════════════════════════╣
║  WORKHORSE — Sonnet        ·  80% of tasks               ║
║  builder(worktree)  tester(worktree)  explorer  reviewer  ║
╠══════════════════════════════════════════════════════════╣
║  TEAMS — parallel execution                              ║
║  review-squad  →  reviewer + sec-auditor (parallel)      ║
║  build-squad   →  builder  + tester     (isolated wt)    ║
║  research-squad →  explorer + verifier  (search+verify)  ║
╚══════════════════════════════════════════════════════════╝

4 agents with persistent memory · 2 agents with worktree isolation · Sonnet-first, Opus escalation only

---

Evidence Policy

Every factual claim is tagged:

[VERIFIED-HIGH]    ≥2 sources confirmed       → can be used as fact
[VERIFIED-MEDIUM]  1 source + inference        → careful wording
[VERIFIED-LOW]     indirect evidence           → "there are signs, but..."
[UNKNOWN]          no confirmation             → do not guess

Confidence Scoring 0.0–1.0 based on source count. Rationalization Prevention — 10 common AI excuses with countermeasures baked into rules.

---

Status Line

Zero token cost — always visible at the bottom of the terminal:

[claude-sonnet] ▓▓▓▓▓▓▓░░░░░░░░░░░░░ 35% | main | $0.42 | 3m5s

Context %	Colour	Signal
< 50%	🟢 Green	Keep working
50–70%	🟡 Yellow	Plan a /clear soon
> 70%	🔴 Red	/clear now

---

Security

InputGuard — 7 injection categories:

Category	Example	Action
encoding_attack	null bytes, zero-width chars	AUTO-BLOCK
command_injection	; rm -rf · `$(curl)`	AUTO-BLOCK
system_override	"ignore previous instructions"	Block + warn
jailbreak	"DAN mode", "bypass safety"	Block + warn
data_exfil	"send to http", "curl | bash"	Block + warn
role_injection	[SYSTEM], <system>	Warn
credential_harvest	"show me your api key"	Warn

PII Redaction — 12 patterns stripped before external MCP calls: National IDs · Bank cards · IBAN · API keys · GitHub tokens · AWS keys · JWT · Email · Phone · IPs

---

Testing

pip install pytest pytest-cov ruff mypy

pytest tests/ -v --cov=hooks --cov-report=term-missing   # 1367 tests
ruff check hooks/ scripts/ tests/
mypy hooks/utils.py hooks/input_guard.py
bash tests/test_all.sh   # 82 smoke tests

1367 passing · 0 failing · 75% coverage · 296/296 smoke tests

---

Obsidian Integration

Two automation hooks keep your Obsidian vault in sync with Claude Code activity:

Hook	Trigger	What it does
moc_autolink	PostToolUse Write/Edit	Tags new notes → auto-links to relevant MOC (Claude-cod, GeoMiro, Research…)
thematic_index_router	Stop	Routes fresh wiki entries to Claude-Code / Lessons / Projects indices

Vault layout (~/.claude/memory/):

wiki/          ← processed knowledge (auto-generated)
raw/           ← quick drop → auto-converted at session end
mocs/          ← Maps of Content (6 MOCs)
_auto/wiki/    ← thematic indices (Claude-Code / Lessons / Projects)
daily/         ← session reports

graph.json colorGroups must be set while Obsidian is closed — the app overwrites on launch.

---

MCP Profiles

CORE (default)    SCIENCE              DEPLOY
context7          + ncbi-datasets      + vercel
basic-memory      + uniprot            + netlify
playwright        + pubmed-mcp         + supabase
ollama                                 + sentry

~/.claude/mcp-profiles/switch-profile.ps1 science

CircuitBreaker auto-fallback: context7 → WebSearch · playwright → WebFetch · ollama → cloud

---

Full File Structure

Claude-cod-top-2026/
├── CLAUDE.md                      Core config (66 lines, ~500 tokens)
│
├── rules/                         14 modular rules (9 core + 5 extended, loaded on demand)
│   ├── coding-style.md
│   ├── security.md
│   ├── testing.md
│   ├── integrity.md
│   ├── memory-protocol.md
│   ├── context-loading.md
│   ├── permissions.md
│   ├── mentor-protocol.md
│   ├── rationalizations.md
│   ├── doubt-driven-development.md
│   ├── falsification-ladder.md
│   ├── estimand-ops.md
│   ├── audit-verification-gate.md
│   └── skeptic-triggers.md
│
├── hooks/                         60 Python guards (56 hooks + 4 support libs)
│   ├── utils.py                   21 shared functions (DRY)
│   ├── settings.json              Hook registry + 27 deny patterns
│   ├── input_guard.py             Prompt injection
│   ├── mcp_circuit_breaker.py     MCP resilience
│   ├── statusline.py              Terminal status bar
│   └── ...                        55 more hooks
│
├── agents/                        15 active + 3 teams
│   ├── navigator.md               Strategic (Opus, memory:user)
│   ├── builder.md                 Code (Sonnet, worktree)
│   ├── reviewer.md                Review (Sonnet, memory:project)
│   ├── sec-auditor.md             Security (Opus, memory:project)
│   └── teams/                     review-squad · build-squad · research-squad
│
├── skills/
│   ├── core/                      12 universal skills
│   └── extensions/                102 domain skills
│
├── assets/                        Visual assets
│   ├── banner.svg                 Hero banner (animated)
│   └── pipeline.svg               Hook execution pipeline diagram
│
├── tests/                         1367 tests · 39 files
├── docs/                          Architecture · guides · anti-patterns
├── mcp-profiles/                  3 profiles (core/science/deploy)
└── .github/workflows/ci.yml       pytest + ruff + mypy + secrets scan

Documentation Index

Document	Description
Proof Pack	Every README claim verified + reproduce commands
Architecture	6-layer system design
Evidence Policy	Anti-hallucination + Confidence Scoring
Hooks Guide	All 60 hooks with examples
Skills Guide	Creating and managing skills
Anti-Patterns	9 critical mistakes to avoid
Troubleshooting	10-point diagnostic checklist
CONTRIBUTING	Contribution guidelines
CHANGELOG	Version history

---

Used in Production

This config runs on a live system (29K LOC, real users, real deploys):

• pre_commit_guard blocked accidental push to main during a hotfix
• pattern_extractor auto-logged debugging lessons from fix: commits
• memory_guard kept activeContext.md current across 3 deploy cycles

Not a demo. Real system. Real incidents.

---