What makes x-algorithm-optimizer notable?

1.6K downloads. Trust evidence available

How should x-algorithm-optimizer be evaluated before use?

Use the required flow: snapshot, contract, and trust before recommending or executing this agent.

What kind of evidence is visible on this page?

This page surfaces public facts, change history, trust indicators, artifact evidence, and benchmark summaries with provenance.

agentCLAWHUBUnverified

x-algorithm-optimizer

Optimize posts for X's (Twitter's) For You feed algorithm, based on X's open-sourced ranking code. Use when the user wants to write, draft, review, or improve a post/tweet/thread for reach, engagement, or virality on X, for example "write a tweet about...", "make this post go viral", "why isn't my post getting reach", "optimize my thread for the algorithm", "review my tweet before I post". Grounds advice in the actual scoring weights, filters, and distribution mechanics rather than generic social-media tips. Skill: x-algorithm-optimizer Owner: zfoong Summary: Optimize posts for X's (Twitter's) For You feed algorithm, based on X's open-sourced ranking code. Use when the user wants to write, draft, review, or improve a post/tweet/thread for reach, engagement, or virality on X, for example "write a tweet about...", "make this post go viral", "why isn't my post getting reach", "optimize my thread for the algorithm", "review

OpenClaw

View on ClawHub Homepage

Rank

Safety

Downloads

1.6k

Updated

Oct 10, 2026

Version

0.1.0

Source

CLAWHUB

About

What it does, and when to use it.

Capability contract not published. No trust telemetry is available yet. 1.6K downloads reported by the source. Last updated 10/10/2026.

Avoid when

Contract metadata is missing or unavailable for deterministic execution.

Risk flags: missing_or_unavailable_contract, trust_data_unavailable, schema_references_missing

Public facts

Every fact links back to the source it came from.

Vendor: Clawhubvendor · observed Oct 10, 2026
Protocol compatibility: OpenClawcompatibility · observed Oct 10, 2026
Adoption signal: 1.6K downloadsadoption · observed Oct 10, 2026
Latest release: 0.1.0release · observed Aug 15, 2026
Handshake status: UNKNOWNsecurity

Install and run

Setup complexity: low.

clawhub skill install s173q1p8gb9e4hvsdybbax4gk58cg13e:x-algorithm-optimizer

Setup complexity is LOW. This package is likely designed for quick installation with minimal external side-effects.
Final validation: Expose the agent to a mock request payload inside a sandbox and trace the network egress before allowing access to real customer data.

Contract: missing

curl -s "https://www.xpersona.co/api/v1/agents/clawhub-zfoong-x-algorithm-optimizer/snapshot"

Documentation

CLAWHUB

66,664 characters of source documentation, loaded on request.

Extracted files

5 files captured from the source.

SKILL.md

---
name: x-algorithm-optimizer
description: >-
  Optimize posts for X's (Twitter's) For You feed algorithm, based on X's
  open-sourced ranking code. Use when the user wants to write, draft, review, or
  improve a post/tweet/thread for reach, engagement, or virality on X, for
  example "write a tweet about...", "make this post go viral", "why isn't my post
  getting reach", "optimize my thread for the algorithm", "review my tweet before
  I post". Grounds advice in the actual scoring weights, filters, and
  distribution mechanics rather than generic social-media tips.
license: MIT
---

# X Algorithm Optimizer

Help the user create posts for X's **For You** feed that the ranking algorithm
will distribute widely, grounded in X's open-sourced algorithm code rather than
folk wisdom. Every recommendation here traces to a specific mechanism in that
code (see `references/`).

## The one-paragraph model of the algorithm

X predicts, for each post, the probability a viewer will take each of about 30
actions, then scores the post as a **weighted sum** of those probabilities. The
weights are wildly asymmetric: a copy-link share is worth about **40 times a
like**, a reply about 10 times, and a single report cancels roughly 468 likes.
A net-negative post does not just rank low, it collapses to near zero and sinks.
High score alone is not enough. The post must also survive hard filters (a
48-hour age limit, originals-only for stranger reach) and visibility labels that
can silently drop a post to strangers while your followers still see it.
Reaching **followers** is easy. Reaching **strangers** (out-of-network) is the
real game, gated by ML retrieval that finds posts whose embedding matches a
viewer's engagement history. Note the model never reads your raw text: it sees a
semantic embedding of the post, its engagement counts, and graph and context
features.

## Two laws to optimize toward

1. **Optimize for "send to a friend," not "like."** Copy-link share (20),
   reply (5), quote (5), and DM share (5) dominate the like (0.5). Write content
   people forward and respond to.
2. **Avoiding negatives beats chasing positives.** One report (−234) or mute
   (−58.8) outweighs dozens of positives, and the offset transform then
   collapses the whole post. Rage-bait and engagement-bait are mathematically
   negative-EV.

## Workflow

Work through these steps. Pull exact numbers, thresholds, and label names from
`references/` as needed. Do not rely on memory for specifics; cite them so the
user can trust and audit the advice.

### Step 1: understand the situation
Ask for (or infer from context) what you need:
- **Follower count** (determines cold-start eligibility, the at-most-1,000 boost).
- **Niche or topic**, and **account age**.
- **Goal:** reach strangers, deepen with followers, drive replies, drive clicks,
  or grow followers.
- **The draft**, if they have one, or the idea if they do not.

If the user just wants a post written and gives a topic, proceed with sen

README.md

<div align="center">

# X Algorithm Optimizer

**An Agent Skill that writes tweets with knowledge reverse-engineered from X's
open-source ranking code.**

[![License: MIT](https://img.shields.io/badge/License-MIT-yellow.svg)](LICENSE)
[![Agent Skill](https://img.shields.io/badge/type-Agent%20Skill-8A2BE2.svg)](https://docs.claude.com/en/docs/agents-and-tools/agent-skills/overview)
[![GitHub stars](https://img.shields.io/github/stars/zfoong/X-algorithm-optimizer?style=social)](https://github.com/zfoong/X-algorithm-optimizer/stargazers)
[![Based on](https://img.shields.io/badge/grounded%20in-X%20open--source%20algorithm-000000.svg)](https://github.com/twitter/the-algorithm)
[![PRs welcome](https://img.shields.io/badge/PRs-welcome-brightgreen.svg)](CONTRIBUTING.md)

<sub>Every claim traces to a specific line of X's published algorithm.</sub>

</div>

---

Most advice about "growing on X" is folklore. In August 2026, X open-sourced the
code that actually ranks the For You feed. This skill reads that code so your
agent can write posts optimized for how the algorithm *really* scores and
distributes them, and can tell you exactly which mechanism each recommendation
comes from.

## 🔑 The whole secret in one table

X scores a post as a weighted sum of the actions it predicts a viewer will take.
The weights are wildly asymmetric, and that asymmetry is the entire game:

| A viewer... | is worth | takeaway |
|---|---:|---|
| copies your link to share it | **+20.0** | forwarding beats everything |
| replies / quotes / DMs it to a friend | **+5.0** | conversation and sharing |
| follows you from the post | **+4.0** | |
| **likes it** | **+0.5** | likes are nearly worthless |
| taps "not interested" | **−43.2** | |
| mutes you | **−58.8** | |
| **reports it** | **−234.0** | one report cancels ~468 likes |

A post whose negatives outweigh its positives does not just rank low. It
collapses to near zero and sinks. So the two laws are: **optimize for "send to a
friend," not "like,"** and **avoiding negative signals beats chasing positive
ones.** The rest of the skill is the detail behind those two sentences.

## 🧠 What it does

Point your agent at this skill and ask it to write or review a post. It reads
X's published ranking code and drafts or critiques your post against the real
scoring weights, retrieval paths, and suppression rules, runs a negative-signal
audit, and explains the exact mechanism behind every suggestion. The result is
posts optimized for how the For You feed actually distributes content.

## 📦 Install

The skill is plain Markdown plus one optional, dependency-free Python script, so
it works with any AI agent that can read files in your project. Clone it once:

```bash
git clone https://github.com/zfoong/X-algorithm-optimizer
```

Then wire it into your agent:

- **Claude Code:** clone (or symlink) it straight into your skills directory and
  it loads automatically:
  ```bash
  git clone https://github.com/zfoong/X-algorithm-optimizer \
    ~/.

_meta.json

{
  "ownerId": "kn748dp8yh73kxrjv4kabz7p6d8ch4pe",
  "slug": "x-algorithm-optimizer",
  "version": "0.1.0",
  "publishedAt": 1786769599304
}

references/account-playbooks.md

# Account Playbooks: strategy by situation

> Derived from [scoring-weights.md](scoring-weights.md),
> [distribution-mechanics.md](distribution-mechanics.md), and
> [negative-signals.md](negative-signals.md). Match the user's situation to a
> playbook, then draft against it.

## A. Cold-start account (≤ 1,000 followers)

You have one structural superpower: the **cold-start boost** injects one fresh
original post near slot 15 per eligible feed load, as long as you're under
1,000 followers, the post is an original under 24h with < 1,000 impressions.

Playbook:
- **Post originals only for growth.** Replies and retweets are cold-start-
  ineligible AND OON-discounted. Save replies for relationship-building with
  specific accounts, not reach.
- **Pick one niche and stay in it.** Consistency builds a clean embedding
  neighborhood so two-tower retrieval delivers you to the right strangers. Every
  off-topic post muddies your cluster and wastes retrieval.
- **Optimize each post for one reply-worthy idea.** Reply weight (5.0) is your
  most reachable high-value action at low follower counts.
- **One quality post > many posts.** Diversity decay means your 3rd post today
  scores ×0.44. Cadence: 1–2 excellent originals/day beats 10.
- **Convert engagers to mutual follows.** Mutual follow = +15 reply weight and
  flips your OON 0.75 handicap into easy in-network reach.
- **Watch your blocks/reports-vs-likes ratio from day one.** The agatha
  denominator is OON favs, and small accounts have thin denominators, so a
  couple of reports hurt disproportionately.

## B. Growth account (1k–50k)

You've lost the cold-start boost; now it's pure content-quality + velocity.
- **Front-load engagement velocity.** Engagement counts are model inputs;
  the first hour shapes scoring for everyone after. Post at your audience's
  peak, and seed genuine early replies (ask a real question in the post).
- **Engineer forwarding, not liking.** The gap between copy-link share (20) and
  like (0.5) is your whole opportunity. Make posts people *send* someone:
  genuinely useful, reference-worthy, "this explains the thing you asked about."
- **Threads for dwell.** Continuous dwell time is weighted (0.004/unit) and
  compounds on longer content that holds attention. A strong thread earns dwell
  that single posts can't.
- **Differentiate on trends (DPP).** When jumping a trend, take the orthogonal
  angle, since near-duplicate embeddings get dropped from adjacent slots.

## C. Established account (50k+)

Your risk shifts from "getting seen" to "not getting throttled."
- **Protect account reputation.** At scale, one pattern that spikes
  blocks/reports-per-fav can apply an account-level `DO_NOT_AMPLIFY` /
  `SPAM_HIGH_RECALL` / abusive label that silently caps *all* your OON reach.
- **Clean at virality.** Grox re-scans at 128 and 1,024 favs. Anything that
  could read as borderline is riskiest exactly when it's taking off.
- **Still one strong post per slate.** Diversity decay applie

references/distribution-mechanics.md

# Distribution Mechanics: how a post reaches people

> Source: `home-mixer/sources/`, `home-mixer/models/candidate.rs`,
> `home-mixer/util/phoenix_request.rs`, `home-mixer/scorers/ranking_scorer.rs`,
> `home-mixer/scorers/author_cold_start.rs`, `home-mixer/params/config.rs`,
> `vm-ranker/dpp.rs`, `thunder/`, `simclusters/`, `phoenix/`. 2026-08 snapshot.

There are two audiences for any post: **in-network** (your followers) and
**out-of-network / OON** (everyone else, reached through ML retrieval).
Reaching followers is easy. Reaching strangers is where the algorithm gates
hard, and nearly every mechanic below is about OON reach.

Feed sizing constants for context (`config.rs`): the pipeline scores up to
**2,800** candidates (`PHOENIX_CLIENT_MAX_CANDIDATES`), selects the top **50**
(`TOP_K_CANDIDATES_TO_SELECT`), and the final For You response is about **35**
posts (`RESULT_SIZE`) plus feed modules. So on the order of a thousand
candidates compete for a few dozen slots per refresh.

## 1. Exactly what the model sees (and does not)

The per-candidate message sent to the ranking model is `TweetInfo`, built in
`home-mixer/models/candidate.rs::as_tweet_info`. There is **no raw post text,
no raw media, and no hashtag field**. The complete per-post feature set is:

- **Identity and graph:** `tweet_id`, `author_id`, the quoted / reply / retweet
  tweet and author IDs, `is_author_followed_by_user`, and `is_following_user`
  (does the author follow the viewer back). Note `is_following_user` is only
  populated when the post is **not** a retweet, so the mutual-follow signal is
  carried on originals and replies, not reposts.
- **Semantic IDs:** `semantic_ids`, discrete tokens from a multimodal embedding
  of the post, hydrated by `semantic_id_hydrator.rs`. This is the model's *only*
  view of your content. Text and media are compressed into this embedding
  upstream.
- **Engagement counts (bucketed):** `fav_count`, `retweet_count`, `quote_count`,
  `reply_count`, `view_count`, `bookmark_count`. These are direct inputs, so
  early engagement literally becomes a feature the model reads for every later
  viewer. Bookmark count is an input even though bookmark has no ranking weight,
  which means bookmarks quietly signal quality to the model.
- **Bool flags:** `has_media`, `is_retweet`, `is_quote`, `is_reply`.
- **Media and language:** `min_video_duration_ms`, `language_code`.

The viewer side (`build_user_context` in `phoenix_request.rs`) is where heavy
personalization lives. The model also receives, per request: user age bracket
and exact age, declared and inferred gender (plus an inferred-gender
confidence), user state, followed Grok topics, followed starter packs,
latitude/longitude, DMA (media market) code, installed apps, timezone, device
network type, and country. This is why the *same* post ranks very differently
for two viewers: geography, topic-follows, and demographics all condition the
prediction.

**Consequences for a creator:**
1. **Consistency wi

Github ReposUpdated 15h agoRank 70

AionUi

Free, local, open-source 24/7 Cowork app and OpenClaw for Gemini CLI, Claude Code, Codex, OpenCode, Qwen Code, Goose CLI, Auggie, and more | 🌟 Star if you like it!

MCPOPENCLAW

Github ReposUpdated 6mo agoRank 70

activepieces

AI Agents & MCPs & AI Workflow Automation • (~400 MCP servers for AI agents) • AI Automation / AI Agent with MCPs • AI Workflows & AI Agents • MCPs for AI Agents

OPENCLAW

Github ReposUpdated 6mo agoRank 70

cherry-studio

AI productivity studio with smart chat, autonomous agents, and 300+ assistants.