# Kodaos CLI

> CLI for managing AI agent skills in the Kodaos ecosystem

- Skill: `kodaos/kodaos-cli` (Agent Skill)
- Install (CLI): `npx skillmds@latest add kodaos/kodaos-cli`
- Raw SKILL.md: https://api.skillmd.com/api/skills/kodaos/kodaos-cli/raw
- Safety review: pending
- Works with: Claude Code, Claude.ai, OpenAI Codex
- Category: AI & ML
- License: MIT
- Author: kodaos (https://skillmd.com/u/kodaos)
- Updated: 2026-09-17
- Page: https://skillmd.com/skills/kodaos/kodaos-cli

---


# kodaos CLI

CLI for managing the AI ecosystem with Kodaos. Provides skills and sessions commands for installation, updates, listing, migration, and session resume workflows.

## Commands

### `kodaos skills add <source>`

Add skills from a GitHub repository.

```
kodaos skills add owner/repo
kodaos skills add https://github.com/owner/repo
```

**Arguments:**

- `<source>` — GitHub shorthand (`owner/repo`) or full GitHub URL

**Options:**

- `-s, --skill <names...>` — Specific skills to install (default: all)
- `-l, --list` — List available skills without installing
- `-y, --yes` — Skip confirmation prompts
- `--copy` — Copy files instead of symlinking
- `-g, --global` — Install to global directory (`~/.kodaos`) instead of project

### `kodaos skills remove [skills...]`

Remove installed skills.

```
kodaos skills remove my-skill
kodaos skills remove --all
```

**Arguments:**

- `[skills...]` — Skills to remove

**Options:**

- `-g, --global` — Remove from global directory
- `-y, --yes` — Skip confirmation prompts
- `--all` — Remove all installed skills

### `kodaos skills list`

List all installed skills.

```
kodaos skills list
kodaos skills list --global
```

**Options:**

- `-g, --global` — List global skills

Displays a table of skill name, source, path, and commit hash.

### `kodaos skills update [skills...]`

Update installed skills to the latest version from GitHub.

```
kodaos skills update
kodaos skills update my-skill --yes
```

**Arguments:**

- `[skills...]` — Skills to update (default: all)

**Options:**

- `-g, --global` — Update global skills
- `-p, --project` — Update project skills
- `-y, --yes` — Skip confirmation prompts

### `kodaos skills migrate [platform]`

Migrate skills from external platforms.

```
kodaos skills migrate vercel
kodaos skills migrate vercel ./skills-lock.json --global
```

**Arguments:**

- `[platform]` — Platform to migrate from (currently supports `vercel`)
- `[file]` — Source lock file path (defaults to `skills-lock.json`)

**Options:**

- `-g, --global` — Write to global lock file
- `-y, --yes` — Skip confirmation prompts

### `kodaos sessions list`

List Claude/Codex sessions with optional source and date filtering.

```
kodaos sessions list
kodaos sessions list --source claude --limit 20
kodaos sessions list --from 2026-05-01 --to 2026-05-10 --output json
```

**Options:**

- `-s, --source <source>` — Session source (`claude`, `codex`, `all`)
- `--from <date>` — Start date filter (`YYYY-MM-DD`)
- `--to <date>` — End date filter (`YYYY-MM-DD`)
- `--limit <number>` — Maximum number of results
- `-o, --output <mode>` — Output mode (`text`, `json`)

### `kodaos sessions resume <id>`

Resume a Claude or Codex session by id.

```
kodaos sessions resume <session-id>
kodaos sessions resume <session-id> --source codex
kodaos sessions resume <session-id> --dry-run
```

**Arguments:**

- `<id>` — Session id to resume

**Options:**

- `-s, --source <source>` — Source override (`claude`, `codex`, `auto`)
- `--dry-run` — Print command preview without executing
- `-o, --output <mode>` — Output mode (`text`, `json`)

## Architecture

- **Lock-driven design** — `kodaos-lock.json` tracks installed skills with source repository and git commit hash
- **Two-tier storage** — Skills are stored in `.agents/skills/` with symlinks created in `.claude/skills/`
- **Git sparse checkout** — Uses `git clone --depth 1 --filter=blob:none --sparse` for efficient cloning
- **Skill discovery** — Skills are discovered via `skills/index.json` (explicit list) or by scanning `skills/*/` directories containing `SKILL.md`
- **Session adapters** — Sessions are aggregated from `~/.claude/history.jsonl` and `~/.codex/session_index.jsonl` via a normalized record model