# Velog Cli

> >

- **Type:** Skill
- **Install:** `agentstack add skill-hamsurang-kit-velog-cli`
- **Verified:** Yes — security-reviewed for prompt injection and unsafe behavior
- **Seller:** [hamsurang](https://agentstack.voostack.com/s/hamsurang)
- **Installs:** 0
- **Category:** [Agent Skills](https://agentstack.voostack.com/c/agent-skills)
- **Latest version:** 0.1.0
- **License:** MIT
- **Upstream author:** [hamsurang](https://github.com/hamsurang)
- **Source:** https://github.com/hamsurang/kit/tree/main/plugins/velog-cli/skills/velog-cli

## Install

```sh
agentstack add skill-hamsurang-kit-velog-cli
```

Requires the [AgentStack CLI](https://agentstack.voostack.com/docs/cli). Works with Claude Code, Cursor, and any MCP-compatible agent.

## About

# velog-cli Skill

`velog` is a CLI client for [velog.io](https://velog.io), a developer blogging platform.
It lets you manage blog posts entirely from the terminal — create from markdown files,
edit, publish drafts, browse trending posts, and more.

## Prerequisites

The `velog` binary must be installed:

```bash
# macOS
brew tap hamsurang/velog-cli && brew install velog-cli

# Cargo (any platform)
cargo install velog-cli
```

If `velog` is not found, tell the user to install it first.

## Authentication

Most write operations require authentication. Check status before attempting mutations.

```bash
# Check if logged in
velog auth status

# Login (interactive — prompts for tokens from browser DevTools)
velog auth login

# Logout
velog auth logout
```

**How to get tokens**: The user opens velog.io in their browser, opens DevTools (F12) →
Application → Cookies, and copies `access_token` and `refresh_token`. The CLI prompts
for these during `velog auth login`.

Credentials are stored at `~/.config/velog-cli/credentials.json` and auto-refresh when expired.

## Post Management

### List Posts

```bash
# List your own posts (requires auth)
velog post list

# List your drafts
velog post list --drafts

# Trending posts (filterable by period)
velog post list --trending
velog post list --trending --period week

# Recent posts across velog
velog post list --recent

# Posts by a specific user
velog post list -u 

# Control result count (1–100, default 20)
velog post list --limit 50

# Pagination
velog post list --recent --cursor 
velog post list --trending --offset 20
```

### Show a Post

```bash
# Show your own post by slug
velog post show 

# Show another user's post
velog post show  -u 
```

The slug is the URL path segment (e.g., `my-first-post` from `velog.io/@user/my-first-post`).

### Create a Post

```bash
# From a markdown file (saved as draft by default)
velog post create -t "Post Title" -f article.md

# From stdin
echo "# Hello World" | velog post create -t "Hello"

# Publish immediately (not draft)
velog post create -t "Title" -f article.md --publish

# With tags and custom slug
velog post create -t "Title" -f article.md --tags "rust,cli,velog" --slug my-custom-slug

# Private post
velog post create -t "Title" -f article.md --publish --private
```

### Edit a Post

```bash
# Update content from a file
velog post edit  -f updated.md

# Update title
velog post edit  -t "New Title"

# Update tags
velog post edit  --tags "new,tags"

# Update everything
velog post edit  -t "New Title" -f updated.md --tags "new,tags"
```

### Publish a Draft

```bash
velog post publish 
```

### Delete a Post

```bash
# With confirmation prompt
velog post delete 

# Skip confirmation
velog post delete  -y
```

## Output Formats

All commands accept `--format` to control output style:

| Format | Use case | Description |
|--------|----------|-------------|
| `pretty` | Human reading (default) | Tables, colors, markdown rendering |
| `compact` | AI agents, scripts | Minified JSON on a single line |
| `silent` | CI/CD, exit-code-only checks | Queries emit JSON, mutations emit nothing |

When Claude is running velog commands programmatically, use `--format compact` to get
structured JSON output that's easy to parse:

```bash
velog post list --format compact
velog post show my-post --format compact
```

## Common Workflows

### Write and publish a blog post

1. Write markdown content to a file
2. Create as draft: `velog post create -t "Title" -f article.md --tags "tag1,tag2"`
3. Preview on velog.io if needed
4. Publish: `velog post publish `

### Update an existing post

1. Find the slug: `velog post list --format compact`
2. Edit: `velog post edit  -f updated.md`

### Browse trending posts for inspiration

```bash
velog post list --trending --period week --limit 10
```

## Shell Completions

```bash
# Generate completions
velog completions zsh    # or bash, fish, elvish, powershell
```

## Error Handling

1. **Not authenticated**: Run `velog auth status` first. If not logged in, guide
   the user through `velog auth login`.
2. **Post not found**: Verify the slug with `velog post list --format compact`.
3. **Binary not found**: Tell the user to install via `brew` or `cargo`.

## Tips for AI Agent Usage

- Always use `--format compact` when you need to parse the output programmatically.
- Check auth status before attempting create/edit/delete/publish operations.
- When creating posts from user-written content, save to a temp file first, then
  pass via `-f`. This avoids shell escaping issues with stdin piping.
- Default behavior is to save as draft — use `--publish` only when the user
  explicitly asks to publish immediately.
- Use `--format silent` for mutations when you only care about success/failure
  (check exit code).

## Source & license

This open-source skill is cataloged on AgentStack and links to its original source — we do not rehost the code.

- **Author:** [hamsurang](https://github.com/hamsurang)
- **Source:** [hamsurang/kit](https://github.com/hamsurang/kit)
- **License:** MIT

Install and usage instructions live in the source repository linked above.

## Pricing

- **Free** — Free

## Security capabilities

Automated source analysis of v0.1.0 — what this tool can access:

- **Network access:** no
- **Filesystem access:** no
- **Shell / process execution:** no
- **Environment & secrets:** no
- **Dynamic code execution:** no

*"Yes" means the capability is present in the source — more access means more to trust, not that it is unsafe.*


## Versions

- **0.1.0** — security scan: passed — Imported from the upstream source.

## Links

- Listing page: https://agentstack.voostack.com/l/skill-hamsurang-kit-velog-cli
- Seller: https://agentstack.voostack.com/s/hamsurang
- Browse the marketplace: https://agentstack.voostack.com/browse

---
Listed on AgentStack — the marketplace for AI agent skills and MCP servers. Every listing is security-reviewed. Creators keep 70%.