Skip to content

relunctance/project-standard

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

27 Commits
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 

Repository files navigation

project-standard

Knowledge Base — Project Inspection Standards

⚠️ This repository contains ONLY knowledge (standards, schemas, reference docs). It has no code. Code implementations live in auto-evolve.


The Nineteen-Perspective Project Inspection Standard Library — the evaluation engine behind auto-evolve.

Every project, nineteen lenses: User → Product → Project → Tech → Market Influence → Business Sustainability → Security → Performance → Testing → Integration → Observability → Documentation → i18n → Accessibility → Reliability → Cost Efficiency → Compatibility → Industry Vertical → Business/Compliance. One consistent standard.

Version: v1.1.0 | Total checks: 116 | License: MIT

中文: README.zh-CN.md


✨ Why project-standard?

"Is this project any good?"
  ↓
Developer A: "Code looks fine"
Developer B: "Features are implemented"
Developer C: "I think it's usable"

No common language → Everyone talks past each other.

project-standard gives project evaluation a shared language — nineteen perspectives + consistent standards = comparable, trackable inspection results.


🎯 Key Features

Nineteen-Perspective Inspection Framework

Not nineteen teams' work — nineteen lenses for one person to examine any project:

Perspective Ask Yourself Finds
👤 User Is it pleasant to use? CLI design, error messages, learning curve
📦 Product Does it deliver what it promises? README vs reality, unresolved pain points
🏗 Project Is it managed well? Learnings loop, inspection rhythm
⚙️ Tech Is the code healthy? Duplicates, complexity, dependencies
📊 Market Influence Is it visible and growing? Stars, ecosystem, competitive edge
💼 Business Sustainability Can it survive long-term? Funding, governance, maintenance plan
🔒 Security Is it vulnerable to attacks? Injection, XSS, auth, secrets, threat modeling
Performance Is it efficient under load? Latency budgets, memory, DB efficiency
🧪 Testing Is it protected from regressions? Coverage, CI, integration, strategy
🔗 Integration Does it integrate cleanly? Dependency health, API contracts
📡 Observability Can you diagnose it when it breaks? Logs, metrics, traces, alerting
📚 Documentation Can users and contributors succeed? Onboarding, reference, architecture
🌍 i18n Can global users use it? Locale, RTL, content localization
Accessibility Can users with disabilities use it? WCAG, keyboard nav, screen reader
🔄 Reliability Does it fail gracefully? SLA, failover, backup, DR
💰 Cost Efficiency Is it cost-effective? Compute, storage, AI/ML, CI/CD costs
🔀 Compatibility Can it evolve without breaking users? API versioning, backward/forward compat
🏢 Industry Vertical Does it meet industry requirements? HIPAA/PCI-DSS/FERPA/FedRAMP/IEC 62443
⚖️ Business/Compliance Is it legally compliant? License compliance, data privacy, IP

Three-Tier Perspective Classification

Tier Perspectives Behavior
Required User, Product, Project, Tech, Security, Testing Always active, cannot be disabled
Type-Required Performance, Integration, Observability, Documentation, i18n, Accessibility, Reliability, Cost Efficiency, Compatibility, Business/Compliance Active based on project type
Optional Market Influence, Business Sustainability, Industry Vertical Off by default, enable explicitly

Quick Scan vs Full Scan

Mode Scans Use When
Quick Required + Type-Required perspectives Daily CI, fast feedback
Full All 19 perspectives Release inspection, comprehensive review

Two-Level Project Taxonomy

Level 1: Business Form (determines which perspectives are active)
  ├── Frontend     → Web / Mobile / Desktop / Plugin / Mini-app
  ├── Backend      → REST API / Microservice / CLI / DevOps / Middleware
  ├── AI/Agent    → Skill / Agent / ML Pipeline / AI Service
  ├── Infrastructure → IoT / Blockchain / Data Pipeline
  ├── Content     → SSG Docs / API Docs / Static Blog
  └── Generic

Level 2: Tech Stack (determines specific check items)
  Python / JavaScript-TS / Go / Rust / Java / PHP / Solidity / ...

📂 Complete Reference Docs

references/
├── user/                                 ← User experience checklist (9 checks)
├── product-requirements.md               ← Product promises vs reality
├── project-inspection.md                 ← Project health checklist
├── project-types.md                      ← Taxonomy + weights + scan modes
├── tech/                                 ← Code quality checklist
├── market-influence/                     ← Community traction + competitive visibility
├── business-sustainability/             ← Funding + governance + longevity
├── security/                             ← Security audit + threat modeling (27 checks)
├── performance/                          ← Performance budgets + efficiency (18 checks)
├── testing/                             ← Test coverage + CI matrix (17 checks)
├── integration/                         ← Dependency health + API contracts
├── observability/                       ← Logs + metrics + tracing + alerting
├── documentation/                       ← Onboarding + reference + architecture
├── i18n/                               ← Locale support + RTL + content localization
├── accessibility/                       ← WCAG + keyboard nav + screen reader
├── reliability/                         ← SLA + failover + backup + DR
├── cost-efficiency/                     ← Compute + storage + AI/ML + CI/CD costs
├── compatibility/                       ← API versioning + backward/forward compat
├── industry-vertical/                   ← Healthcare/Finance/Education/Gov/IoT (45 checks)
├── business-compliance/                 ← License compliance + data privacy + IP + regulatory
│
├── scoring-algorithm.md                  ← Unified scoring methodology
├── arbitration-rules.md                  ← Perspective boundaries + conflict resolution
├── perspective-dependencies.md           ← Scanner execution order + data flow
├── version-migration.md                 ← Version evolution + breaking change policy
├── user-interaction-protocol.md         ← Tier 1-3 confirmation flow
├── perspective-config-schema.md          ← perspective-config.yaml complete schema
├── perspective-config.example.yaml       ← Annotated config example
├── project-types-examples/              ← Per-type config templates
│   ├── ai-agent.example.yaml
│   ├── backend.example.yaml
│   └── frontend.example.yaml
├── check-registry/                      ← All 116 check IDs (source of truth)
├── fix-action-registry/                 ← Standardized fix action vocabulary
├── scanner-contract/                    ← Scanner interface specification
├── report-generator/                    ← Report format specs
├── scan-history/                        ← Storage format + trend definition
├── auto-evolve-integration.md          ← How auto-evolve loads and executes
│
├── code-standards.md                   ← Code quality + naming conventions
├── architecture.md                    ← Architecture standards + review template
├── security.md                        ← Security guidelines
├── git-workflow.md                    ← Git workflow + PR template
├── quality.md                         ← Test tiers + 100% coverage requirement
├── risk-management.md                 ← Risk management + Pre-mortem method
├── product-design.md                  ← Product design + Jobs anti-entropy principles
├── metrics.md                         ← Product metrics system
└── inspection-template.md             ← Inspection report template

Framework docs (at root):

├── README.md                          ← This file
├── README.zh-CN.md                    ← 中文版
├── CHANGELOG.md                      ← Version history
├── VERSION                           ← Current version (1.1.0)
├── QUICK-REFERENCE.md                ← One-page summary card
└── CONTRIBUTING.md                   ← Contributor quick start guide

🚀 Quick Start

With auto-evolve (recommended)

clawhub install project-standard
clawhub install auto-evolve

# Done — auto-evolve auto-loads the standard library
python3 scripts/auto-evolve.py scan --dry-run

Standalone as inspection reference

# Read a perspective checklist
cat references/user/user-perspective.md

# See all check statistics
python3 scripts/check_stats.py

# Validate everything
python3 scripts/validate_check_ids.py
python3 scripts/validate_links.py
python3 scripts/validate_registry.py

📊 Check Statistics (v1.1.0)

Total: 116 checks across 19 perspectives

By perspective:     SEC 27 | PERF 18 | TEST 17 | USR 9 | IND 45
By severity:        Critical 11 | High 34 | Medium 23 | Low 3
Auto-actionable:    27 (38%) | Human judgment: 44 (62%)

🔍 Project Type Weight Reference

See references/project-types.md for the complete weight matrix.

Business Form Required Type-Required Weight Range
Frontend 5 always on a11y 15%, docs 10%, perf 10%
Backend 5 always on reliability 20%, compat 20%, sec 20%
AI/Agent 5 always on cost 25%, perf 10%, testing 15%
Infrastructure 5 always on sec 30%, reliability 25%, cost 20%
Content/Docs 3 of 5 docs 40%, i18n 10%, a11y 5%
Generic 5 always on balanced across all

💡 Signature Check Items

Pre-mortem Risk Management

"Assume the project already failed — why?" This finds 30% more problems than traditional risk assessment.

┌─────────────────────────────────────┐
│  👥 PEOPLE  │ ⚙️ PROCESS │ 💻 TECH │ 🌍 EXTERNAL  │
│  Key person left  │  Aggressive timeline  │  Scaling failed   │  Market shift   │
└─────────────────────────────────────┘

Jobs Anti-Entropy Product Principles

"Don't ask users what they want — they don't know until you show them."

  • Focus is saying no (cut 350 ideas, ship 10)
  • Redefine the category (not first to market, but best execution)
  • Perfect the invisible (good wood on the cabinet back)

Test Tiers + 100% Coverage Requirement

Tier Ratio Covers
Unit 70% Core logic, utility functions
Integration 20% Cross-module interactions
E2E 10% Core user paths

Core logic (Logic/Service/Dao) requires 100% coverage.


🏗 Relationship with auto-evolve

project-standard  ← Defines standards (referee, read-only knowledge base)
       ↑
       │ 19 perspectives + 116 checks + arbitration rules
       │
auto-evolve      ← Executes inspection (athlete, executable code)
       ↓
  findings → user confirmation → closed-loop improvement

Clear boundary: project-standard has no code, no scanners, no LLM calls. auto-evolve has all the execution logic. They never cross.


Validation Scripts

python3 scripts/validate_check_ids.py    # Detect duplicate IDs, format validation
python3 scripts/validate_links.py        # Check markdown internal links
python3 scripts/validate_registry.py     # Validate registry table completeness
python3 scripts/check_stats.py           # Check statistics + distribution
python3 scripts/check_alignment.py       # Registry ↔ perspective doc alignment

Related Projects

  • auto-evolve — Nineteen-perspective automated inspection engine
  • SoulForce — AI Agent memory evolution system
  • hawk-bridge — OpenClaw context memory integration

License

MIT

About

The Four-Perspective Project Inspection Standard Library — evaluation criteria for auto-evolve.

Topics

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

 
 
 

Contributors

Languages