fix(seo): repoint fork metadata and install docs, add SEO/GEO positioning - #6
Conversation
…ale keywords All five manifests carried upstream URLs (EveryInc repo, every.to article) and keywords for removed capabilities (image-generation, rails, ruby). Repository/homepage/websiteURL now point at the fork; keywords/tags reflect the current 16-skill/16-agent inventory. Release-owned version and description fields untouched; release:validate and bun test green. Op: correct Restores: spec:manifest-metadata-describes-this-fork
…d of upstream npm The documented bunx @every-env/compound-plugin invocations resolved via npm to upstream's published CLI, installing the 39-skill upstream product instead of this 16-skill fork. Point install and cleanup commands at bunx github:OutlineDriven/compound-engineering-slim, which resolves this fork's bundled plugins, and swap the npm version badge for a GitHub release badge. Op: correct Restores: spec:documented-install-installs-this-fork
Correct the stale skill count from 14 to 16 and add a Why this fork section comparing this slimmed fork against upstream EveryInc/compound-engineering-plugin. The section includes a dated comparison table, slimming facts, and a standalone-quotable FAQ answer for answer engines. Op: extend
Replace the Getting Started section with a Quickstart that walks from install through /ce-setup to a first /ce-code-review in about five minutes, with prerequisites up front and next-step links to the skill inventory, per-skill docs, and other-platform installs. Op: extend
Add an llms.txt index at the repo root in llmstxt.org shape so language models can discover the plugin's purpose, install commands, doc set, and the 16-skill/16-agent inventory of this slimmed fork. Op: extend
Add a durable runbook documenting the audit/optimize/probe/iterate loop for the fork's public discovery surfaces. It enumerates owned surfaces, target queries, the read-only probe procedure, an empty baseline table for probe runs to fill, and the iteration cadence tied to inventory changes and releases. Op: extend
Probed 10 query/engine pairs on 2026-06-07, recording 20 baseline rows. Op: extend
Durable ce-plan artifact for the seo-geo-loop branch; execution tracked by the commits on this branch. Op: extend
|
You have reached your Codex usage limits for code reviews. You can see your limits in the Codex usage dashboard. |
![gemini-code-assist[bot]](https://avatars.githubusercontent.com/in/956858?s=60&v=4){kind=small}
There was a problem hiding this comment.
Code Review
This pull request implements an SEO and GEO optimization loop for the compound-engineering-slim fork. It repoints manifest metadata (repository, homepage, and tags) to the new fork, adds a detailed 'Why this fork' comparison section and a 5-minute quickstart to the READMEs, routes installation commands through the fork instead of upstream npm, and introduces an llms.txt file and an SEO/GEO runbook. A review comment correctly identifies a mathematical inconsistency in the README regarding the number of skills removed during slimming (39 - 16 = 23, not 38).
Important
The consumer version of Gemini Code Assist on GitHub is being sunset. Starting June 18, 2026, new organization installations will be blocked, and all code review activity will officially cease on July 17, 2026.
For more details on the timeline and next steps, please review the Help Documentation.
| | converter targets | 5 (opencode, codex, pi, gemini, kiro) | more (copilot and droid dropped here) | | ||
| | bundled plugins | 1 | 2 (coding-tutor removed here) | | ||
|
|
||
| - 38 skills were removed during slimming. |
There was a problem hiding this comment.
There is a mathematical inconsistency between the comparison table and this bullet point. The table states that this fork has 16 skills and upstream has 39 skills. Therefore, 23 skills were removed during slimming (39 - 16 = 23), not 38.
| - 38 skills were removed during slimming. | |
| - 23 skills were removed during slimming. |
There was a problem hiding this comment.
Pull request overview
This PR updates the fork’s public-facing SEO/GEO surfaces so install instructions and metadata consistently point to OutlineDriven/compound-engineering-slim, and adds supporting documentation artifacts (GEO inventory + loop runbook) to improve factual discoverability.
Changes:
- Repoints manifest metadata (homepage/repository/tags/keywords) from upstream to this fork across package/plugin/marketplace files.
- Updates root and plugin READMEs with fork positioning, corrected component counts (16 skills / 16 agents), and routes
bunxinstall/cleanup commands through the GitHub fork rather than upstream npm. - Adds
llms.txt, an SEO/GEO loop runbook, and a dated execution plan for maintaining/probing discoverability.
Reviewed changes
Copilot reviewed 10 out of 10 changed files in this pull request and generated 2 comments.
Show a summary per file
| File | Description |
|---|---|
README.md |
Adds “Why this fork” positioning, updates counts, replaces npm badge, and repoints bunx commands to the GitHub fork. |
plugins/compound-engineering/README.md |
Adds a 5-minute quickstart and links to inventory/docs. |
plugins/compound-engineering/.cursor-plugin/plugin.json |
Repoints homepage/repository and refreshes keywords for the fork. |
plugins/compound-engineering/.codex-plugin/plugin.json |
Repoints homepage/repository/websiteURL and refreshes keywords for the fork. |
plugins/compound-engineering/.claude-plugin/plugin.json |
Repoints homepage/repository and refreshes keywords for the fork. |
package.json |
Repoints homepage/repository and updates package description text. |
llms.txt |
Adds a GEO-oriented inventory + install summary artifact. |
docs/seo-geo-loop.md |
Adds a runbook for maintaining and probing SEO/GEO surfaces, including a baseline table. |
docs/plans/2026-06-07-001-feat-seo-geo-loop-plan.md |
Adds an execution plan documenting intent/constraints for the SEO/GEO campaign. |
.claude-plugin/marketplace.json |
Repoints plugin homepage and refreshes marketplace tags for the fork. |
💡 Add Copilot custom instructions for smarter, more guided reviews. Learn how to get started.
| | converter targets | 5 (opencode, codex, pi, gemini, kiro) | more (copilot and droid dropped here) | | ||
| | bundled plugins | 1 | 2 (coding-tutor removed here) | | ||
|
|
||
| - 38 skills were removed during slimming. |
| "name": "@every-env/compound-plugin", | ||
| "version": "3.11.2", | ||
| "description": "Official Compound Engineering plugin for Claude Code, Codex, and more", | ||
| "description": "Slim fork of the Compound Engineering plugin for Claude Code, Codex, and more", |
…arison table The "38 skills removed / 92+ personas consolidated" bullets counted cumulative cleanup-registry entries (renames, merges, intermediate states), which a reader subtracts from the current-state table (16 vs 39 / 16 vs 43) and gets a contradiction. The table alone carries the differentiation; the registry-sweep bullet keeps the upgrade-hygiene story without subtraction bait. Op: correct Restores: spec:comparison-numbers-internally-coherent
Summary
SEO + GEO campaign for the fork's public surfaces. Plan:
docs/plans/2026-06-07-001-feat-seo-geo-loop-plan.md.Fixes (wrong data):
package.json, plugin.json x3, marketplace.json) pointedhomepage/repositoryat upstream and carried keywords for removed capabilities (image-generation,rails,ruby) — repointed to this fork, keywords refreshed. Release-ownedversion/descriptionuntouched.bunx @every-env/compound-plugininstall/cleanup commands in the README resolved via npm to upstream's CLI (installing the 39-skill upstream product). Now route throughbunx github:OutlineDriven/compound-engineering-slim— verified end-to-end against a scratchCODEX_HOME(16 fork agents installed, zero upstream components). npm badge (upstream's version) replaced with a GitHub release badge.Positioning (new content, claims fact-backed only):
llms.txt(GEO artifact) with verified inventory names.docs/seo-geo-loop.mdrunbook: owned surfaces, target queries, probe procedure, iteration cadence, and a 20-row GEO baseline (fork currently absent from all target queries; upstream ranks refactor(plugin): slim compound-engineering — remove 9 skills, 6 agents, merge 3 skill pairs, deslop prose #1-2 — the measuring stick for iteration 2).Also done (outside the diff): GitHub About description + 10 topics set via
gh repo edit.Verification
bun test: 885 pass / 0 failbun run release:validate: in sync (16 agents, 16 skills)Op:trailer; no breaking markersDeferred
package.jsonstill carries upstream's name@every-env/compound-plugin(no publish workflow exists). Renaming/publishing the fork's CLI under its own npm name is a separate product decision.