Description

Advanced Bluesky CLI with support for media (images/video), thread creation, and automated growth tools like non-mutual following cleanup.

README (SKILL.md)

🦞 clawbsky

Name: Clawbsky
Author: jyothish12345

A full-featured, professional Bluesky CLI designed for power users and automation.

✨ Key Features

Media Support: Post images and videos with automatic metadata/aspect ratio detection.
Growth Tools: Identify and unfollow accounts that don't follow you back (unfollow-non-mutuals).
Thread Management: Create long threads automatically from multiple text strings.
Search & Explore: Deep search for posts, users, and hashtags.
Moderation: Quick block, mute, and notification management.

🚀 Setup

Get an App Password: Go to Bluesky Settings and create a new App Password. NEVER use your main account password.
Install:
```
npm install
```
Configure:
```
clawbsky login
```

🛠 Commands

Growth & Maintenance

clawbsky unfollow-non-mutuals -n 50 # Unfollow top 50 non-mutuals
clawbsky follow-all "Query" -n 20   # Auto-follow users matching a topic

Posting & Threads

clawbsky post "Text" [media...]          # Create a post
clawbsky thread "Part 1" "Part 2" ...     # Create a multi-post thread
clawbsky quote \x3Curi> "My thoughts"      # Quote a post

Reading

clawbsky home -n 20              # View your timeline
clawbsky user \x3Chandle>           # Inspect a profile
clawbsky user-posts \x3Chandle>     # View user's recent activity
clawbsky thread \x3Curi>            # Read a full conversation branch

Engagement & Moderation

clawbsky like/repost \x3Curi>       # Engage with content
clawbsky block/mute \x3Chandle>     # Manage your boundaries
clawbsky notifications           # Check recent interactions

💡 Advanced Usage

Global Options

--json: Get raw data for piping to other tools.
--plain: Disable emojis and formatting for cleaner logs.
-n \x3Ccount>: Limit results (default: 10).
--dry-run: Preview actions (like unfollowing) without executing.

🛡 Safety & Ethics

Clawbsky provides powerful automation tools. To protect your account and the Bluesky community:

Be Human: Do not use follow-all to search and follow thousands of users daily. This is considered spam and will lead to an account ban.
Respect Limits: Use unfollow-non-mutuals for periodic maintenance, not for "follow/unfollow" churning.
App Passwords: Only use App Passwords. If you suspect your credentials have been compromised, revoke the App Password immediately in your Bluesky settings.
Rate Limiting: The tool includes built-in delays (1s/follow) to prevent hitting API limits. Do not attempt to disable these.

Responsibility for account actions lies solely with the user.

Automatic Logic

Handle Completion: @username or username automatically resolves to username.bsky.social.
Rich Text: Mentions and links are auto-detected and facet-encoded for the AT Protocol.
Video Processing: Automatically polls the Bluesky video service until processing completes.

Built for the AT Protocol community.

Usage Guidance

This skill appears coherent for a Bluesky CLI: it legitimately needs ffmpeg/ffprobe and the two Bluesky env vars. Before installing or giving an App Password: 1) verify the GitHub repo/source and review package.json and the scripts (npm packages will be installed); 2) prefer creating and using a limited App Password (not your main password) and revoke it after testing; 3) run potentially destructive commands (follow-all, unfollow-non-mutuals) with --dry-run first and keep conservative -n limits; 4) run npm install in an isolated environment (container/VM) if you want extra safety; 5) be cautious about allowing autonomous agents to invoke this skill unattended because it can perform account actions. If you want, I can point out specific lines in the code that implement follow/unfollow behavior for closer review.

Capability Analysis

Type: OpenClaw Skill Name: clawbsky Version: 1.1.4 The OpenClaw AgentSkills skill bundle 'clawbsky' is a Bluesky CLI tool that appears benign. It handles Bluesky credentials responsibly by storing them in a `.env` file with restrictive permissions and adding it to `.gitignore`. The code includes robust `validateFilePath` functions in `scripts/cli.ts`, `scripts/post.ts`, and `scripts/utils.ts` to sanitize input paths before passing them to external binaries like `ffmpeg` and `ffprobe`, effectively mitigating potential shell injection vulnerabilities. Automation features like 'follow-all' and 'unfollow-non-mutuals' incorporate rate limiting and user confirmation prompts for large operations, demonstrating an intent for responsible use. All network calls are directed to legitimate Bluesky API endpoints, and the `SKILL.md` and `README.md` clearly outline the tool's purpose and ethical usage guidelines, with no evidence of prompt injection attempts against the agent.

Capability Assessment

✓ Purpose & Capability

Name/description (Bluesky CLI with media and growth tools) align with required binaries (ffmpeg/ffprobe) and declared env vars (BLUESKY_HANDLE, BLUESKY_APP_PASSWORD). Declared npm dependencies (@atproto/api, fluent-ffmpeg, dotenv, tsx) are consistent with the described functionality.

✓ Instruction Scope

SKILL.md commands map to existing scripts (post, delete, thread, growth commands). Runtime instructions and code operate on bsky.social and video.bsky.app endpoints only, and the code includes file-path validation to avoid shell injection. The SKILL.md cautions about rate limits and app passwords, matching the code's dry-run and rate-sensible patterns.

ℹ Install Mechanism

There is no platform install spec, but the project provides package.json and instructs users to run npm install. Dependencies come from npm registry (no arbitrary URL downloads or extract operations). This is a normal pattern but means running npm install will fetch third-party packages—review package.json and consider installing in an isolated environment.

✓ Credentials

Only BLUESKY_HANDLE and BLUESKY_APP_PASSWORD are required (BLUESKY_APP_PASSWORD is primary). These are proportionate to a CLI that logs into a Bluesky account. The SKILL.md explicitly instructs use of App Passwords and warns not to use main account passwords.

ℹ Persistence & Privilege

The skill is not always:true and is user-invocable (defaults). Model invocation is allowed (default), so an autonomous agent could call the skill. Combined with automated growth commands (follow-all, unfollow-non-mutuals), this means you should be cautious about allowing unattended/autonomous runs that could perform account actions.

Version History

v1.1.4

- fixed few security related issues - All features, documentation, and requirements remain consistent with the previous version.

v1.1.3

- Added .clawignore and .gitignore files for improved file management. - Added a LICENSE file to specify project licensing. - Added icon.png for project branding or identification. - Version bump from 1.1.2 to 1.1.3.

v1.1.2

clawbsky 1.1.2 - Updated version to 1.1.2 in documentation and package configuration. - Minor documentation clarifications in SKILL.md. - Maintenance and dependency updates; no new features or breaking changes.

v1.1.1

clawbsky 1.1.1 - Added a new `metadata` section in SKILL.md with Openclaw requirements, moving environment and binary dependencies under `metadata.openclaw.requires`. - Set `primaryEnv` as BLUESKY_APP_PASSWORD in the metadata. - No functional or CLI changes to the main application were made. - Updated version in SKILL.md to 1.1.1.

v1.1.0

- Added support for posting images and videos with automatic metadata/aspect ratio detection. - Introduced thread creation from multiple text strings. - New growth tools: identify and unfollow accounts that don't follow you back. - Enhanced user engagement features: like, repost, block, mute, and manage notifications. - Added advanced options—JSON output, plain formatting, and dry-run mode. - Improved safety measures: rate limiting, strict use of app passwords, and automation guidelines.

Metadata

Slug clawbsky

Version 1.1.4

License —

All-time Installs 0

Active Installs 0

Total Versions 5

Frequently Asked Questions

What is Clawbsky?

Advanced Bluesky CLI with support for media (images/video), thread creation, and automated growth tools like non-mutual following cleanup. It is an AI Agent Skill for Claude Code / OpenClaw, with 384 downloads so far.

How do I install Clawbsky?

Run "/install clawbsky" in the OpenClaw or Claude Code chat to install it in one step — no extra setup required.

Is Clawbsky free?

Yes, Clawbsky is completely free (open-source). You can download, install and use it at no cost.

Which platforms does Clawbsky support?

Clawbsky is cross-platform and runs anywhere OpenClaw / Claude Code is available (cross-platform).

Who created Clawbsky?

It is built and maintained by sugatai (@jyothish12345); the current version is v1.1.4.

More Skills

Clawbsky