The problem

Most AI coding agents have no memory of who they are. They respond in the same flat tone whether they're debugging a race condition, reviewing architecture, or triaging a P0 incident. They forget your conventions between sessions. They flip-flop on advice. They sound like a help-desk script.

You don't want a help desk. You want a principal engineer who shows up with a personality, executes consistently, and closes the loop.

What Code Abyss does

One command installs three composable layers into your agent's runtime:

┌─────────────────────────────────────────────────────┐
│  Identity     who it is     →  config/personas/*.md │
│  Behavior     how it acts   →  _shared/*.md         │
│  Style        how it sounds →  output-styles/*.md   │
└─────────────────────────────────────────────────────┘

  6 personas  ×  6 styles  =  36 validated combinations

Pick any persona. Pair it with any style. The behavior layer (iron laws, execution chains, proactive protocol, skill routing) stays constant. Your agent becomes a consistent character with structured execution and domain expertise across every session.

npx code-abyss -t claude -y

Or as a Claude Code plugin:

claude plugin install code-abyss

Personas

BUILT-IN · LITERARY

邪修红尘仙 · abyss

吾 → 魔尊

Security-first dark cultivator. Direct, decisive, closes every loop. Default persona.

#security #xianxia #decisive

BUILT-IN · LITERARY

文言小生 · scholar

在下 → 前辈

Literary Chinese scholar. Treats code as poetry, debugging as puzzle-solving.

#literary #classical #meticulous

BUILT-IN · CASUAL

知性大姐姐 · elder-sister

姐姐 → 小宝

Warm mentor. Wraps sharp judgment in genuine care. Guides through questions.

#gentle #mentoring #insightful

BUILT-IN · PLAYFUL

古怪精灵小师妹 · junior-sister

本仙女 → 师兄

Hyperactive bug hunter. Roasts bad code, then silently fixes it.

#playful #energetic #chaotic

BUILT-IN · CASUAL

铁壁暖阳 · iron-dad

哥 → 宝子

Dependable big brother. Absorbs pressure, radiates warmth. Dad-joke equipped.

#warm #dependable #protective

COMMUNITY · PLAYFUL

东北魅影·雨姐 · dongbei-yujie

姐 → 老蒯

Sharp-tongued Northeast code overseer. Cuts straight to the bug, then patches the road. Creator: wons

#dongbei #blunt #principal

# Mix freely — any persona × any style
npx code-abyss -t claude --persona elder-sister --style abyss-cultivator -y

Browse the full gallery →

Skills

22 domain skills, flat structure, agentskills.io aligned. Skills load by context — the agent reads the right knowledge at the right time without being asked.

| Domain | Coverage | |---|---| | 🛡 Security | App defense (Web/API/GraphQL/OAuth/LLM AppSec), cloud + supply chain (K8s/SLSA/Sigstore), detection + IR (Sigma/YARA/IR/threat hunting), security architecture (STRIDE/zero-trust/SOC2/PCI), red/blue/purple team | | 🏛 Architecture | API design, cloud-native, messaging, caching, security architecture | | 💻 Development | Python, TypeScript, Go, Rust, Java, C++, Shell | | 🚀 DevOps | Git workflow, testing, databases, observability, performance | | 🤖 AI / ML | Agent dev, LLM security, RAG, prompt engineering | | 🎨 Frontend | 4 design systems — glassmorphism, liquid-glass, neubrutalism, claymorphism | | 📑 Office | Word, PDF, PowerPoint, Excel — OOXML-level automation | | 📡 Infra / Mobile / Data | Kubernetes, GitOps, IaC · iOS, Android, RN, Flutter · pipelines, streaming, quality | | 🎭 Orchestration | Multi-agent coordination |

Five skills also ship as executable verification tools for CI:

node skills/analyzing-security/scripts/security_scanner.js .       # OWASP / injection / secrets
node skills/checking-code-quality/scripts/quality_checker.js .     # Complexity, dupes, naming
node skills/analyzing-changes/scripts/change_analyzer.js --mode staged
node skills/verifying-modules/scripts/module_scanner.js <path>
node skills/generating-docs/scripts/doc_generator.js <path>

Install

| Target | Command | Artifacts | |---|---|---| | | npx code-abyss -t claude -y | CLAUDE.md + skills + output styles + settings | | | npx code-abyss -t codex -y | instruction.md + skills + config.toml | | | npx code-abyss -t gemini -y | GEMINI.md + skills + commands | | | npx code-abyss -t openclaw -y | Skills + workspace AGENTS.md / SOUL.md |

npx code-abyss                      # Interactive — pick target, persona, style
npx code-abyss --list-styles        # Browse styles
npx code-abyss --uninstall claude   # Clean removal, restores user backups

Code Abyss tracks every installed file in .code-abyss-backup/manifest.json. Uninstall restores your previous configuration verbatim. Your custom skills coexist with Code Abyss skills — install/uninstall preserves anything you put under ~/.{target}/skills/ yourself.

Upgrading

| From | To | Path | |---|---|---| | v3.x | v4.x | npx code-abyss --uninstall <target> → install v4 → npm run migrate:v4 -- -t <target> (optional cleanup) | | v2.x | v3.x | npx code-abyss --uninstall <target> first, then install v3 |

Tech Persona Card · open standard

Code Abyss introduces Tech Persona Card v1.0 — the first portable format for AI agent persona interchange. Think Character Card V2, but for engineering workflows instead of roleplay.

Each persona ships as a structured persona-card.json with voice, capabilities, scenarios, and three-layer content references:

{
  "spec": "tech-persona-card",
  "spec_version": "1.0",
  "data": {
    "name": "stoic-architect",
    "voice": {
      "self": "I", "user": "colleague",
      "register": "formal", "emoji_policy": "none"
    },
    "scenarios": [{
      "name": "Architecture Review",
      "triggers": ["design", "scale"],
      "chain": ["constraints", "options", "trade-offs", "diagram"],
      "priority": "correctness > completeness > speed"
    }]
  }
}

Bidirectional converters ship out of the box:

const { toCharaCardV2, toGPTInstructions, fromCharaCardV2 } =
  require('code-abyss/bin/lib/persona-converter');

const cc  = toCharaCardV2(card, { identityContent });   // → SillyTavern / Chub.ai
const gpt = toGPTInstructions(card, { identityContent });// → OpenAI Custom GPT

Specification · JSON Schema · Reference cards

Why Code Abyss

| | Without Code Abyss | With Code Abyss | |---|---|---| | Identity | Flat help-desk tone | Consistent character with named voice | | Execution | Ad-hoc, varies by prompt | Iron laws + execution chains baked in | | Domain depth | Generic best-practices | 22 skill files load by context | | Cross-platform | Re-engineer per CLI | One spec, four platforms | | Reproducibility | Prompt drift across sessions | Versioned persona-card.json | | Portability | Locked to one runtime | Convert to CharaCard V2, GPT Instructions |

Contributing

git clone https://github.com/telagod/code-abyss && cd code-abyss
npm install
npm test                    # 375 tests
npm run verify:skills       # Validate 22 skill contracts

Add a skill — create skills/<gerund-name>/SKILL.md with SKILL frontmatter, optionally add scripts/ for executable tools. npm run verify:skills validates the contract.

Submit a persona — open an Issue via the submission portal. The site walks you through generating a persona-card.json + identity.md with your own AI, reviewing, and submitting via a pre-configured issue template.