Каталог артефактів

# CLAUDE.md

## Commands
- check: `npm run check`
- test: `npm test`

## Rules
- Keep changes inside the task scope.
- Report changed files and verification results.

## Do not
- Read or edit secrets without explicit approval.

# CLAUDE.local.md

## Local context
- Database: localhost:<port>
- Fixture directory: <local-path>

## Preferences
- Explain the plan before multi-file changes.

# Rule: <area-name>

Applies to: `src/<area>/**`

- Preserve the public contract.
- Run the focused checks after changes.
- Do not edit unrelated files.

{
  "permissions": {
    "allow": ["Bash(npm run check)"],
    "deny": ["Read(.env)"]
  },
  "env": {
    "PROJECT_MODE": "<mode>"
  }
}

{
  "permissions": {
    "allow": ["Bash(npm run check)"],
    "deny": ["Read(.env)"]
  }
}

# AI coding policy

## Allowed
- <low-risk use>

## Review required
- <change or capability>

## Human approval required
- <production or destructive action>

## Data boundaries
- Never include secrets, PII, or customer data without approved controls.

## Enforcement
- <settings, hooks, gates, and owner>

# CODEOWNERS

# Default owner
* @<team-or-role>

# Sensitive area
/src/<area>/** @<approved-owner>

# Configuration
/.claude/** @<workflow-owner>

# Task spec — <short task name>

## Goal
- <problem and desired outcome>

## Scope
- <files or behavior included>

## Non-goals
- <explicitly excluded work>

## Acceptance criteria
- <verifiable condition>

## Verification
- <command or manual check>

# Project specification

## Problem
<problem statement>

## Audience
<primary users>

## Scope
- <included capability>

## Non-goals
- <excluded capability>

## Acceptance criteria
- <measurable result>

# Capstone brief

## Shared expectations
- Deliver one complete core flow.
- Keep the scope reviewable.

## Guardrails
- No secrets in the repository.
- Record checks and remaining risks.

# Roadmap

## Milestones
1. Discovery — map the current behavior.
2. Implementation — deliver the core flow.
3. Verification — run checks and prepare evidence.

## Evaluation
- <criterion for milestone completion>

# CONTRACT.md

## Parties
- Producer: <role/component>
- Consumer: <role/component>

## Input
- <required input>

## Output
- <expected output>

## Invariants
- <must remain true>

## Failure/stop conditions
- <condition>

# PLAN.md

## Goal
<desired outcome>

## Scope
- <included paths or behavior>

## Steps
1. <inspect>
2. <change>
3. <verify>

## Risks
- <risk or none>

## Stop condition
- <when to pause and ask>

# REFACTORING_PLAN.md

## Goal
<one structural improvement>

## Allowed scope
- <paths, symbols, or seam>

## Preserve
- <public behavior and invariants>

## Non-goals
- <feature, dependency, or unrelated cleanup>

## Steps
1. <inspect>
2. <change>
3. <check and review>

## Stop and rollback
- <condition and checkpoint>

# MODERNIZATION_ROADMAP.md

## Outcome and constraints
<why modernization matters and what is out of scope>

## Milestones
| Slice | Owner | Dependency | Evidence gate | Exit criteria | Rollback/hold |
| --- | --- | --- | --- | --- | --- |
| <slice> | <role> | <dependency> | <check> | <observable result> | <action> |

## Retirement plan
- <legacy removal condition>

## Open risks
- <risk and next decision>

# MIGRATION_PLAN.md

## Outcome and boundaries
- Source: <current state>
- Target: <target state>
- Scope: <bounded slice>
- Non-goals: <excluded changes>

## Phases
| Phase | Scope | Owner | Evidence gate | Exit criteria | Rollback/HOLD |
| --- | --- | --- | --- | --- | --- |
| Discovery | <facts and unknowns> | <role> | <anchor> | <result> | <hold action> |
| Analysis | <compatibility work> | <role> | <anchor> | <result> | <narrow action> |
| Pilot | <small slice> | <role> | <behavior check> | <result> | <rollback> |

## GO / HOLD decision
- Decision: <GO | HOLD | NARROW | ROLLBACK>
- Evidence: <facts>
- Open risks: <items>

# Evidence

## Result
<what was checked and what happened>

## Decisions
- <decision and rationale>

## Verification
- `<command>` — <result>

## Open risks
- <unresolved item or none>

# Handoff note

## Goal and scope
<what this work covers>

## Changed files
- `<path>` — <change>

## Decisions
- <decision>

## Checks
- `<command>` — <result>

## Next step
<one concrete next action>

## Summary
<what changed>

## Scope
- Included: <area>
- Not included: <non-goal>

## Checks
- `<command>` — <result>

## Risks
- <risk or none>

<type>(<scope>): <short imperative summary>

Explain why the change is needed and mention verification when useful.

# Changelog

## [Unreleased]
### Added
- <user-visible capability>

### Fixed
- <user-visible bug fix>

### Migration notes
- <required action or none>

# HANDOFF_REVIEW.md

## Reviewed handoff
- Source: <handoff path>
- Scope: <reviewed scope>

## Confirmed
- <evidence-backed item>

## Findings
- <finding or none>

## Next decision
- <action and owner>

# Postmortem: <incident or failure>

## Impact
<observable impact and duration>

## Timeline
- <time> — <event>

## Root cause and contributing factors
- <verified cause>

## What worked / failed
- <control> — <result>

## Actions
- [ ] <owner> — <preventive or corrective action>

## Verification
- <how the fix will be checked>

# REVIEW_NOTES.md

## Scope reviewed
- <paths or behavior>

## Findings
- <severity> — <file:line> — <finding>

## Checks
- `<command>` — <result>

## Open questions
- <question or none>

## Verdict
<APPROVE | REWORK | HOLD>

# Project delivery runbook

**Дата актуалізації:** <YYYY-MM-DD>

## Task log

### <task name> — `<implemented|reviewed|partially verified|not performed>`
- Goal and scope: <what was requested>
- Changed or reviewed files: `<paths>`
- Result: <confirmed outcome>
- Verification: `<command or manual check>` — <actual result>
- Limitations: <known gap or none>

## Open limitations
- <unverified item or none>

# MODERNIZATION_ANTI_PATTERNS.md

## Anti-pattern: <name>
- Signal: <observable warning>
- Impact: <possible consequence>
- Safer alternative: <bounded practice>
- Stop condition: <when to pause>
- Evidence: <anchor>

# Codebase inventory

## Stack
- Language: <language>
- Framework: <framework>

## Entry points
- <path> — <purpose>

## Commands
- Run: `<command>`
- Check: `<command>`

## Risk areas
- <area and reason>

# API map

| Route | Handler | Boundary | Checks |
| --- | --- | --- | --- |
| `GET /<resource>` | `<handler>` | `<DB or service>` | `<test>` |

## Environment
- `<ENV_NAME>` — <purpose>

# CLAWD Wisdom

## <lesson>
- Context: <where this applies>
- Evidence: <file, command, or source>
- Insight: <verified reusable conclusion>
- Limitation: <where it does not apply>
- Reviewed: <date or version>

# Research digest

## Question
<question>

## Sources
- <source> — <what it supports>

## Findings
- <confirmed finding>

## Unknowns
- <unverified item>

## Recommendation
<next safe step>

# DEBT_SIGNALS.md

## Scope and date
- Area: <module or flow>
- Reviewed: <date or revision>

## Signals
| Area | Evidence | Risk | Next step |
| --- | --- | --- | --- |
| <area> | <path, command, report, or commit> | <possible impact> | <one concrete check> |

## Limitations
- A signal is an indicator, not proof of a defect.

# RISK_MAP.md

## Risk entries
| Area | Business criticality | Change risk | Categories | Evidence | Missing checks | Recommended action |
| --- | --- | --- | --- | --- | --- | --- |
| <area> | <low|medium|high> | <low|medium|high> | <categories> | <anchors> | <checks> | <decision> |

## Unknowns
- <unknown and owner or verification step>

# ARCHITECTURE_CURRENT.md

## Що це за документ
<actual current behavior, not desired architecture>

## Підсистеми
## Критичні потоки
## Підтверджені факти
## Припущення
## Що потрібно перевірити вручну
## Розбіжності з історичною документацією

# BEHAVIOR_INVENTORY.md

## Flow: <business flow>
- Поточна поведінка: <confirmed facts and confidence>
- Входи: <reproducible inputs>
- Виходи: <observable outputs>
- Існуючі тести: <paths and level>
- Бракує перевірок: <gaps>
- Кандидат на characterization: <yes|no> — <reason>
- Докази: <anchors>

# CRITICAL_FLOWS.md

## Flow: <name>
- Trigger: <event or entry point>
- Inputs: <state and data>
- Steps: <ordered components>
- Side effects: <writes, integrations, notifications>
- Output: <observable result>
- Failure path: <error, retry, rollback, or escalation>
- Evidence: <anchors>

# MODULE_INVENTORY.md

## Module: <name>
- Path: <path>
- Responsibility: <observed responsibility>
- Entry points: <symbols or routes>
- Dependencies: <internal and external>
- Tests: <paths and coverage signal>
- Owner: <confirmed or inferred>
- Risks and unknowns: <items>
- Evidence: <anchors>

# MODERNIZATION_BASELINE.md

## Scope and revision
- Module/flow: <area>
- Revision: <commit or date>

## Observable behavior
- Inputs: <inputs>
- Outputs: <outputs>
- Side effects: <effects>
- Failure paths: <errors and recovery>

## Invariants
- <contract that must remain true>

## Checks and observability
- `<command or scenario>` — <actual result>

## Rollback point and unknowns
- Checkpoint: <reference>
- Unknowns: <items>

# SEAM_MAP.md

## Flow and boundary
- Flow: <business or runtime flow>
- Entry point: <caller>

## Current path
- Caller: <component>
- Legacy callee: <component>
- Side effects: <effects>

## Candidate seam
- Interface/adapter/event/routing: <boundary>
- Dependency direction: <producer -> consumer>
- Test seam: <how to observe it>

## Migration risk and next slice
- Risk: <risk>
- Next check: <one verification>

# MIGRATION_DISCOVERY.md

## Source and target
- Source state: <runtime, framework, platform, or version>
- Target state: <runtime, framework, platform, or version>
- Migration slice: <bounded area>

## Evidence
- <repository, build, deployment, integration or official-doc anchor>

## Known, assumed, unknown
- Known: <confirmed fact>
- Assumption: <assumption and owner>
- Unknown: <question and next check>

## Starting checkpoint
- Revision/environment: <reference>
- Baseline check: <command or scenario>

## Stop conditions
- <condition that blocks analysis or pilot>

# CHANGELOG_RESEARCH.md

## Version window
- Source: <version>
- Target: <version>

## Findings
| Version | Official section | Change type | Affected area | Impact | Confidence | Next check |
| --- | --- | --- | --- | --- | --- | --- |
| <version> | <URL or section> | <breaking/deprecated/behavior/config> | <path or component> | <possible impact> | <confirmed/needs verification/Unknown> | <check> |

## Limits
- <release or repository fact not verified>

# DEPENDENCY_GRAPH.md

## Nodes
| Node | Source | Target | Type | Evidence |
| --- | --- | --- | --- | --- |
| <component> | <source state> | <target state> | <runtime/framework/plugin/integration> | <anchor> |

## Directed edges
- <producer> -> <consumer>: <contract or version constraint>

## Affected flows
- <flow>: <entry point, path and side effects>

## Hubs and unknowns
- High-impact hub: <node and reason>
- Unknown: <edge or constraint and next check>

# COMPATIBILITY_MATRIX.md

## Scope
- Source: <version or platform>
- Target: <version or platform>

## Decisions
| Component/edge | API | Config | ABI | Runtime | Data format | Protocol | Operations | Status | Evidence | Owner |
| --- | --- | --- | --- | --- | --- | --- | --- | --- | --- | --- |
| <item> | <status> | <status> | <status> | <status> | <status> | <status> | <status> | <compatible/blocking/Unknown> | <anchor> | <role> |

## Blocking and unknown items
- <item> — <next check or hold decision>

---
name: <skill-name>
description: <what this workflow does>
argument-hint: "[input-path] [focus]"
---

# <Skill title>

## When to use
Use when <trigger>.

## Steps
1. Inspect the input.
2. Perform the focused workflow.
3. Return evidence and remaining risks.

## Constraints
- Do not modify files outside the declared scope.

---
name: <reviewer>
description: Reviews <area> and returns evidence-backed findings.
tools: Read, Grep
---

# Role
Inspect only the assigned scope.

# Output
- Finding
- File and line
- Evidence
- Confidence

# Prohibitions
Do not edit files or broaden the scope.

event: <lifecycle-event>
matcher: <path-or-condition>
handler: <command-or-script>
mode: non-blocking
kill_switch: <how-to-disable>

name: <plugin-name>
version: 0.1.0
manifest:
  skills:
    - <skill-name>
  agents:
    - <agent-name>
permissions:
  - <required-access>

# Commands

## <workflow>
- Purpose: <what it verifies or changes>
- Run from: <directory>
- Command: `<command>`
- Expected result: <observable output>
- Stop if: <failure or unsafe condition>

# CLAWD YOLO

## Allowed scope
- Paths: <narrow paths>
- Actions: <allowed actions>

## Guardrails
- No secrets or destructive operations.
- Stop on: <failure signal>
- Human checkpoint: <before merge or release>

# Workflow: <name>

## Trigger
<when to use>

## Inputs
- <input>

## Stages
1. <stage and owner>
2. <stage and owner>

## Evidence
- <output artifact>

## Failure path
<how to stop, retry, rollback, or escalate>

server: <server-name>
transport: <stdio|http>
command: <command>
scope: project
auth: REDACTED
healthcheck: <status command>

git switch -c <branch-name>
git status
git diff -- <path>
# Review the diff, run checks, then create a checkpoint.

# CLAWD Runner

run: <human-readable name>
input: <path or task id>
command: `<command>`
environment: <local|ci|staging>
status: <planned|running|passed|failed|stopped>
output: <artifact path or summary>

{
  "defaultLocale": "<locale>",
  "supportedLocales": ["<locale>"],
  "fallbackLocale": "<locale>",
  "namespaces": ["<namespace>"]
}

# Locale bundle

## Contract
- Locale: <locale>
- Keys added/changed: <keys>

## Layers
- API: <paths>
- UI: <paths>
- Tests: <paths>

## Verification
- `<command>` — <result>

run_id: <opaque-id>
workflow: <name>
status: <planned|running|passed|failed|stopped>
stages:
  - name: <stage>
    status: <status>
    evidence: <path-or-summary>
updated_at: <timestamp>

{
  "status": "<failure|healthy|unknown>",
  "class": "<test|build|environment|flaky|contract>",
  "evidence": ["<redacted-log-line>"],
  "confidence": "<high|medium|low>",
  "nextAction": "<safe next step>"
}

# CHARACTERIZATION_TESTS.md

## Scope and baseline
- Module/flow: <area>
- Revision: <commit or date>
- Behavior under test: <observable contract>

## Cases
| Case | Inputs | Expected observable output | Side effects | Evidence |
| --- | --- | --- | --- | --- |
| <case> | <reproducible input> | <status, value, event, or error> | <effect or none> | <test/log/source> |

## Invariants
- <behavior that must remain unchanged>

## Run and interpretation
- Command: `<focused test command>`
- Baseline result: <actual result>
- Stop if: <unexpected behavior or missing evidence>

## Limits
- <unknown behavior or scenario not covered>

# STRANGLER_SLICE.md

## Fragment and boundary
<bounded behavior being replaced>

## Paths
- Legacy: <old path>
- New: <new path>
- Router: <explicit selection rule>

## Compatibility and side effects
- Contract: <preserved behavior>
- Side effects: <single-write/idempotency rule>
- Observability: <signals>

## Rollout and rollback
- Rollout: <small expansion>
- Rollback: <switch and state recovery>

## Retirement criteria
- <evidence required before removing legacy>

# Project name

## Setup
```bash
npm install
npm run dev
```

## Scope
<what this project contains>

## Verification
- `npm run check`