# codestable/CodeStable

CodeStable 是一套面向严肃软件工程、坚持人在环的 AI 编码工作流。它不以编排 Agent 为中心，而是组织需求、架构、特性、问题与历史决策，让 Codex 和 Claude 驱动的开发过程可控、可追溯、可持续演进。

Repository: https://github.com/codestable/CodeStable
Canonical: https://ross.abutalabs.com/products/codestable
Homepage: https://codestable.dev
Language: Python
License Family: other
Last push: 2026-08-18T01:06:33+00:00

## Health v2 (maintenance only)
Score: 80/100 (v2, computed 2026-09-03T02:20:16.233290+00:00)
- activity 98, release rhythm 96, longevity 10
- inputs: {"age_days": 143, "days_push": 16, "days_rel": 31, "gap_med": 8, "n_releases_24m": 4}
- flags: young, no_license
- formula: round(0.45*activity + 0.35*rhythm + 0.20*longevity); archived -> min(score, 10)

## Adoption (not part of the score)
Stars 1112, forks 83 (observed 2026-08-28T04:03:38.006636+00:00)

## What it is
CodeStable is a set of lightweight skill contracts (installed as Codex/Claude plugins) that structure AI-assisted coding around requirements, evidence, and project memory rather than agent orchestration. It provides skills like /cs-feat, /cs-issue, /cs-refactor, and /cs-review to keep Codex- and Claude-driven development bounded, verifiable, and traceable in long-lived projects.

## Use cases
- manage ai coding workflow for long-term projects
- add guardrails to claude code and codex sessions
- keep project memory across ai coding sessions
- require evidence before bug fixes and refactors
- route feature work and issues through structured skills
- run read-only ai code reviews
- onboard a repository for ai-assisted development

## When to choose
- you use Codex or Claude Code for serious software engineering and want human-in-the-loop control
- you need traceable, evidence-based AI-driven changes in a long-lived codebase
- you want lightweight skill contracts instead of heavyweight agent orchestration frameworks

## When to avoid
- you want fully autonomous multi-agent orchestration
- you need a licensed product for compliance-sensitive environments (no license file is published)
- you don't use Claude Code, Codex, or a skills-compatible CLI

## Facets
- artifact type: plugin
- maturity: active
- function: agent-framework, prompt-engineering, developer-tools, workflow-automation, code-review, testing
- domain: developer-tools, large-language-models, version-control
- platform: cli, cross-platform, python
- tags: ai-coding-workflow, claude-code, codex, skills, human-in-the-loop, project-memory, sdlc, prompt-contracts, automation, ai-agents

## Member repositories
- codestable/CodeStable (main) score 80

## Provenance
- Observed fields: from GitHub, fetched 2026-08-28T04:03:38.006636+00:00.
- Health v2: computed from the inputs above; adoption is never an input.
- Inferred fields (summary, facets, guidance): AI-extracted, prompt v1, taxonomy v1, on 2026-08-30T06:42:42.376559+00:00, confidence not recorded.
  - readme: https://github.com/codestable/CodeStable (fetched 2026-08-28T04:03:38.006636+00:00, sha c91179beaf32)
  - homepage: https://codestable.dev (fetched 2026-08-29T12:46:42.738142+00:00, sha d902298c257e)
- Data as of 2026-08-30T08:39:29.467469+00:00.