[](https://github.com/ChloeVPin/x-algorithm-skill-kit/actions/workflows/validate.yml) [](LICENSE) [](https://github.com/ChloeVPin/x-algorithm-skill-kit/releases/latest) [](https://github.com/ChloeVPin/x-algorithm-skill-kit/stargazers) [](#install) [](https://platform.claude.com/docs/en/agents-and-tools/agent-skills/overview)
Install with the Vercel skills CLI:
npx skills add ChloeVPin/x-algorithm-skill-kitIf you do not have Node.js, clone the repository and copy the skills/x-algorithm folder into your agent's skills directory:
git clone https://github.com/ChloeVPin/x-algorithm-skill-kit.gitx-algorithm-skill-kit packages a self-contained Codex skill that helps AI agents draft, review, score, and optimize X posts based on engineering principles from the public X For You recommendation release.
It provides actionable guidance on retrieval, neural ranking, content filtering, and post mechanics without requiring a clone of the full upstream source tree.
- Post Strategy & Drafting: Structure hooks, formatting, media, and audience targeting for higher For You engagement.
- Algorithm Architecture Insights: Explain recommendation mechanics across Home Mixer, Phoenix, Thunder, and Grox.
- Agent Skill Distribution: Instant installation into Codex environments via
npx.
- Codex environment configured to load skills from
$CODEX_HOME/skills - Node.js version 18 or newer with
npx
- Inspect the bundled skill package:
npx x-algorithm-skill-kit help - Install the skill directly into Codex:
npx x-algorithm-skill-kit install
- Restart your Codex environment so the new skill is registered.
Trigger the skill in your agent conversations to optimize posts or analyze recommendation mechanics:
Use x-algorithm to review this X post draft and optimize its hook for For You retrieval.
Use x-algorithm to generate three post variants tailored for developers interested in AI recommendation systems.
Use x-algorithm to explain how Phoenix retrieval scoring influences post distribution.
- Home Mixer: Timeline orchestration, candidate hydration, and scoring assembly.
- Phoenix Engine: Out-of-network graph processing, neural candidate retrieval, and candidate ranking.
- Thunder Serving: High-speed in-network graph serving and candidate filtering.
- Grox AI: Multimodal content understanding, topic detection, and post quality scoring.
- Pipeline Components: Candidate filters, hydrators, scorers, selectors, ad insertion, brand safety signals, and media engagement metrics.
| File / Path | Purpose |
|---|---|
skills/x-algorithm/SKILL.md |
Primary Codex skill entry point and prompt triggers. |
skills/x-algorithm/references/post-strategy.md |
Actionable post formatting, hook design, and engagement strategy rules. |
skills/x-algorithm/references/overview.md |
High-level architectural summary of the X recommendation stack. |
skills/x-algorithm/references/pipeline.md |
Step-by-step candidate flow from retrieval to final timeline output. |
skills/x-algorithm/references/subsystems.md |
In-depth breakdown of Home Mixer, Phoenix, Thunder, and Grox. |
skills/x-algorithm/references/operations.md |
Operational guidelines, signal weights, and system filters. |
scripts/install-skill.mjs |
Skill installation script for Codex environments. |
docs/release-notes.md |
Detailed version release notes and update history. |
AGENTS.md |
Agent configuration and instructions. |
- Skill Core: Primary instructions and trigger definition.
- Post Strategy Guide: Rules for post drafting and review.
- Architecture Overview: Recommendation stack overview.
- Pipeline Execution: Detailed candidate pipeline lifecycle.
Contributions to improve reference documentation or skill triggers are welcome. Please keep reference files structured cleanly with single-level deep links from SKILL.md.
- License: Skill package licensed under MIT License.
- Project Repository: ChloeVPin/x-algorithm-skill-kit