{"id":50512378,"url":"https://github.com/iret77/x-post-optimizer","last_synced_at":"2026-06-02T21:02:51.221Z","repository":{"id":358070264,"uuid":"1239876156","full_name":"iret77/x-post-optimizer","owner":"iret77","description":null,"archived":false,"fork":false,"pushed_at":"2026-05-15T15:28:07.000Z","size":43,"stargazers_count":1,"open_issues_count":0,"forks_count":0,"subscribers_count":0,"default_branch":"main","last_synced_at":"2026-05-15T16:33:23.414Z","etag":null,"topics":["algorithm","claude-skill","content-creation","recommendation-system","x-twitter"],"latest_commit_sha":null,"homepage":null,"language":null,"has_issues":true,"has_wiki":null,"has_pages":null,"mirror_url":null,"source_name":null,"license":"mit","status":null,"scm":"git","pull_requests_enabled":true,"icon_url":"https://github.com/iret77.png","metadata":{"files":{"readme":"README.md","changelog":null,"contributing":null,"funding":null,"license":"LICENSE","code_of_conduct":null,"threat_model":null,"audit":null,"citation":null,"codeowners":null,"security":null,"support":null,"governance":null,"roadmap":null,"authors":null,"dei":null,"publiccode":null,"codemeta":null,"zenodo":null,"notice":null,"maintainers":null,"copyright":null,"agents":null,"dco":null,"cla":null}},"created_at":"2026-05-15T14:32:42.000Z","updated_at":"2026-05-15T15:46:37.000Z","dependencies_parsed_at":null,"dependency_job_id":null,"html_url":"https://github.com/iret77/x-post-optimizer","commit_stats":null,"previous_names":["iret77/x-post-optimizer"],"tags_count":3,"template":false,"template_full_name":null,"purl":"pkg:github/iret77/x-post-optimizer","repository_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/iret77%2Fx-post-optimizer","tags_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/iret77%2Fx-post-optimizer/tags","releases_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/iret77%2Fx-post-optimizer/releases","manifests_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/iret77%2Fx-post-optimizer/manifests","owner_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/owners/iret77","download_url":"https://codeload.github.com/iret77/x-post-optimizer/tar.gz/refs/heads/main","sbom_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/iret77%2Fx-post-optimizer/sbom","scorecard":null,"host":{"name":"GitHub","url":"https://github.com","kind":"github","repositories_count":286080680,"owners_count":33835766,"icon_url":"https://github.com/github.png","version":null,"created_at":"2022-05-30T11:31:42.601Z","updated_at":"2026-05-26T15:22:16.424Z","status":"online","status_checked_at":"2026-06-02T02:00:07.132Z","response_time":109,"last_error":null,"robots_txt_status":"success","robots_txt_updated_at":"2025-07-24T06:49:26.215Z","robots_txt_url":"https://github.com/robots.txt","online":true,"can_crawl_api":true,"host_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub","repositories_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories","repository_names_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repository_names","owners_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/owners"}},"keywords":["algorithm","claude-skill","content-creation","recommendation-system","x-twitter"],"created_at":"2026-06-02T21:02:48.444Z","updated_at":"2026-06-02T21:02:51.204Z","avatar_url":"https://github.com/iret77.png","language":null,"funding_links":[],"categories":[],"sub_categories":[],"readme":"# x-post-optimizer\n\nA Claude skill that drafts and reviews posts for X (Twitter) based on the open-source [xai-org/x-algorithm](https://github.com/xai-org/x-algorithm) repository (released 2026).\n\nUnlike most \"X algorithm\" guides floating around, this skill is built around a strict epistemic rule: every recommendation is labeled as one of three categories.\n\n- **[FACT]** — directly verifiable from the xai-org/x-algorithm source code or README\n- **[INFERENCE]** — logically derivable from facts, but not stated outright in the repo\n- **[HEURISTIC]** — established creator practice, not in the code; plausible but unverified\n\nSpecifically, the skill **refuses to quote concrete weight multipliers** like \"Retweet = 20×\" or \"Reply = 13.5×\" — those numbers are not in the 2026 open-source release. The `params` module containing actual weights was deliberately excluded by xAI for security reasons. Numbers you see online almost always trace back to the 2023 `twitter/the-algorithm` repo (a different system) or to guesswork.\n\n## What it does\n\nTwo modes:\n\n- **Generate** — give it a topic or rough idea, get drafts in any X format (short post, long post, thread, reply, quote) with an algorithmic trace explaining which engagement signals each design choice targets.\n- **Review** — give it an existing draft, get a structured review against an algorithm-derived checklist with concrete rewrite suggestions.\n\nTriggers on requests like \"write a tweet\", \"review my X post\", \"draft a thread on X\", \"will this perform in the For You feed\", \"score this tweet\", and similar.\n\n## What it knows (and what it doesn't)\n\nGrounded in the published xai-org/x-algorithm code:\n\n- The 15 engagement signals Phoenix predicts (favorite, reply, repost, quote, click, profile_click, video_view, photo_expand, share, dwell, follow_author, not_interested, block_author, mute_author, report)\n- Plus fine-grained signals from `weighted_scorer.rs` (share_via_dm, share_via_copy_link, vqv_score, quoted_click, dwell_time)\n- The 10 pre-scoring filters and 2 post-selection filters\n- Architecture: Thunder (in-network) + Phoenix Retrieval (out-of-network) → Phoenix transformer ranker with candidate isolation → Author Diversity scoring\n\nNot grounded — flagged as `[HEURISTIC]` if mentioned:\n\n- Concrete weight values (excluded from open-source release)\n- Sentiment-based suppression (not in the repo)\n- \"Links hurt reach\" (P(click) is a positive signal; the link tradeoff is inferential)\n- Time-of-day, post-frequency, verified-account effects (not addressed by the open-source ranker)\n\n## Install\n\n### Easiest: download the `.skill` file and drop it into your Claude app\n\nWorks in **every Claude app** — macOS, Windows, Linux, iOS, Android.\n\n1. Download [`x-post-optimizer.skill`](https://github.com/iret77/x-post-optimizer/releases/latest/download/x-post-optimizer.skill) from the [latest release](https://github.com/iret77/x-post-optimizer/releases/latest).\n2. Open \u003chttps://claude.ai/customize/skills\u003e (or click \"Customize\" in the Claude app).\n3. Click the \"+\" next to Skills → \"Upload skill\" → choose the `.skill` file.\n4. Start a new conversation. The skill triggers automatically on relevant prompts.\n\nThat's it. No CLI, no git, no manual file copying.\n\n### Claude Code (CLI users)\n\nFor Claude Code, clone the repo into your skills directory:\n\n```bash\ngit clone https://github.com/iret77/x-post-optimizer.git ~/.claude/skills/x-post-optimizer\n```\n\nRestart your Claude Code session.\n\n### Verify the install worked\n\nAfter restarting, ask Claude in a new conversation:\n\n\u003e Do you have a skill called x-post-optimizer available?\n\nA successful install gets you a \"yes, here's what it does\"; a failed install gets you a \"no, but I can help you draft tweets anyway\". If it failed, open an [issue](https://github.com/iret77/x-post-optimizer/issues).\n\n## Structure\n\n```\nx-post-optimizer/\n├── SKILL.md                       # Workflow + triggers (the manifest Claude reads)\n├── LICENSE                        # MIT\n├── README.md                      # This file\n└── references/                    # Loaded on demand when the skill triggers\n    ├── algorithm-facts.md         # Canonical [FACT] list with repo citations\n    ├── format-playbooks.md        # Per-format guidance (short, long, thread, reply, quote)\n    ├── review-checklist.md        # Six-section review checklist\n    └── example-traces.md          # Two worked examples\n```\n\nThe repo is the skill — the whole folder is what Claude loads. The `.skill` distribution file is just a zip of this same folder.\n\n## Source attribution\n\nPrimary source: [xai-org/x-algorithm](https://github.com/xai-org/x-algorithm), licensed under Apache 2.0. This skill quotes architectural facts and signal names from the README and source files — facts are not copyrightable, and the skill contains no code or substantial text passages from the upstream repository.\n\nFor legacy context only: [twitter/the-algorithm](https://github.com/twitter/the-algorithm) (2023 release, different system).\n\n## Disclaimer\n\nThis skill is **not affiliated with, endorsed by, or sponsored by X Corp, xAI, or any subsidiary**. \"X\", \"Twitter\", \"Phoenix\", \"Thunder\", and \"Grok\" are trademarks of their respective owners and are used here in a nominative, descriptive sense only.\n\nAlgorithmic performance is probabilistic. The skill helps reason about *which* signals a post can target and *why*, based on the published architecture — but it cannot predict whether any specific post will perform, because (a) concrete weights are non-public, (b) the ranker is conditioned on user-specific engagement history, and (c) the underlying transformer is stochastic over its hash-based embeddings.\n\n## License\n\n[MIT](LICENSE) — Copyright (c) 2026 iret77.\n\n## Contributing\n\nIssues and pull requests welcome. Particularly valuable contributions:\n\n- New `[FACT]` entries verifiable against later updates of the upstream repo\n- Counterexamples where a `[HEURISTIC]` is empirically wrong (with evidence)\n- Additional format playbooks (e.g., for X Spaces transcripts, X Articles)\n\nPlease preserve the epistemic-labeling discipline — that's the whole point of the skill.\n","project_url":"https://awesome.ecosyste.ms/api/v1/projects/github.com%2Firet77%2Fx-post-optimizer","html_url":"https://awesome.ecosyste.ms/projects/github.com%2Firet77%2Fx-post-optimizer","lists_url":"https://awesome.ecosyste.ms/api/v1/projects/github.com%2Firet77%2Fx-post-optimizer/lists"}