Rewrite README: outcome-led headline, how it works, works with any agent#57
Rewrite README: outcome-led headline, how it works, works with any agent#57greynewell merged 1 commit intomainfrom
Conversation
- Lead with outcome: "Your AI agent knows your entire codebase — not just the files it's looking at" instead of feature description - Add "How it works" section: 3-step flow from watch → graph files → ask - Add "Works with any AI agent" table: Claude Code, Cursor, Copilot, Windsurf, Aider, and any file-reading agent - Add README badge users can add to their own repos - Move Claude Code integration into the agent table instead of its own major section - Elevate `supermodel setup` in Quick start (was buried) - Keep all command reference and configuration content Closes #53 #54 Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
WalkthroughREADME.md was restructured to lead with outcomes and a "How it works" section instead of feature-focused badges. Installation instructions were updated to reflect Linux/macOS curl setup with a terminal-aware setup wizard. Command examples changed from Changes
Estimated code review effort🎯 2 (Simple) | ⏱️ ~8 minutes Possibly related PRs
Suggested reviewers
Poem
🚥 Pre-merge checks | ✅ 4 | ❌ 1❌ Failed checks (1 warning)
✅ Passed checks (4 passed)
✏️ Tip: You can configure your own custom pre-merge checks in the settings. ✨ Finishing Touches🧪 Generate unit tests (beta)
Comment |
There was a problem hiding this comment.
Actionable comments posted: 2
🧹 Nitpick comments (2)
README.md (2)
5-5: Tiny readability polish: reduce repeated sentence starts.Line 5 has three consecutive sentences starting with “No…”. It reads a bit robotic; a quick rephrase would flow better.
🤖 Prompt for AI Agents
Verify each finding against the current code and only fix it if needed. In `@README.md` at line 5, The three short sentences "No prompt changes. No extra context windows. No new tools to learn." are repetitive; change them into a single, smoother phrase to improve flow—e.g., replace that fragment on line 5 with something like "No prompt changes, extra context windows, or new tools to learn." in the README.md to avoid repeated sentence starts and improve readability.
23-24: Wording is too absolute vs the setup table.Line 23 says “No configuration needed on the agent side,” but Line 41 (Aider
--read) and Line 44+ (Claude live-update hook) show there are practical setup nuances. I’d soften this to “minimal/no special integration required” to avoid confusion.Proposed wording tweak
-`.graph` files are plain text. Any agent that can read files — Claude Code, Cursor, Copilot, Windsurf — picks them up automatically through its normal file-reading tools. No configuration needed on the agent side. +`.graph` files are plain text. Most agents that can read files — Claude Code, Cursor, Copilot, Windsurf — can use them through normal file-reading tools, with minor agent-specific setup in some cases.Also applies to: 33-43
🤖 Prompt for AI Agents
Verify each finding against the current code and only fix it if needed. In `@README.md` around lines 23 - 24, Change the absolute claim in the README sentence ".graph files are plain text. Any agent that can read files — Claude Code, Cursor, Copilot, Windsurf — picks them up automatically through its normal file-reading tools. No configuration needed on the agent side." to a softer phrasing such as "minimal/no special integration required" (or similar) and apply the same tone adjustment to the related passages around lines 33–43 that mention Aider `--read` and the Claude live-update hook so the documentation acknowledges there may be small setup nuances while still conveying low friction.
🤖 Prompt for all review comments with AI agents
Verify each finding against the current code and only fix it if needed.
Inline comments:
In `@README.md`:
- Line 198: Update the website link in README.md to use HTTPS instead of HTTP:
replace the string "http://supermodeltools.com" with
"https://supermodeltools.com" so the table row "| **Website** |
[supermodeltools.com](http://supermodeltools.com) |" consistently uses HTTPS and
matches the rest of the documentation.
- Around line 3-13: The README hero is missing the demo visual requested in
issue `#53`; add a short terminal GIF or screenshot showing `supermodel watch`
output plus the agent reading `.graph` files to demonstrate the flow — embed it
in the hero section right after the rewritten headline (the block starting at
line 3), include alt text and a one-line caption like "Live demo: supermodel
watch + agent using .graph files", and ensure the image file (GIF/PNG) is added
to the repo and referenced by its filename so the README renders the demo.
---
Nitpick comments:
In `@README.md`:
- Line 5: The three short sentences "No prompt changes. No extra context
windows. No new tools to learn." are repetitive; change them into a single,
smoother phrase to improve flow—e.g., replace that fragment on line 5 with
something like "No prompt changes, extra context windows, or new tools to
learn." in the README.md to avoid repeated sentence starts and improve
readability.
- Around line 23-24: Change the absolute claim in the README sentence ".graph
files are plain text. Any agent that can read files — Claude Code, Cursor,
Copilot, Windsurf — picks them up automatically through its normal file-reading
tools. No configuration needed on the agent side." to a softer phrasing such as
"minimal/no special integration required" (or similar) and apply the same tone
adjustment to the related passages around lines 33–43 that mention Aider
`--read` and the Claude live-update hook so the documentation acknowledges there
may be small setup nuances while still conveying low friction.
🪄 Autofix (Beta)
Fix all unresolved CodeRabbit comments on this PR:
- Push a commit to this branch (recommended)
- Create a new PR with the fixes
| Your AI agent knows your entire codebase — not just the files it's looking at. | ||
|
|
||
| Give your AI coding agent a map of your codebase. | ||
| Supermodel maps every file, function, and call relationship in your repo and writes a `.graph` file next to each source file. Your agent reads them automatically via grep and cat. No prompt changes. No extra context windows. No new tools to learn. | ||
|
|
||
| Supermodel CLI connects AI coding agents to the [Supermodel API](https://api.supermodeltools.com), providing call graphs, dependency graphs, dead code detection, and blast radius analysis as context during your sessions. It operates in two modes: **file mode**, which writes `.graph.*` sidecar files next to your source so agents can read them at any time, and **on-demand analysis**, which runs targeted queries against the graph without touching the filesystem. | ||
| ```bash | ||
| curl -fsSL https://supermodeltools.com/install.sh | sh | ||
| ``` | ||
|
|
||
| --- | ||
|
|
||
| ## Links | ||
| ## How it works |
There was a problem hiding this comment.
Missing demo visual in the hero flow (objective miss).
Line 3 onward rewrites the headline well, but the README still lacks the demo GIF/screenshot requested in issue #53. Adding a short terminal capture (supermodel watch stats + agent using .graph data) would close that objective and likely improve conversion.
🧰 Tools
🪛 LanguageTool
[style] ~5-~5: Three successive sentences begin with the same word. Consider rewording the sentence or use a thesaurus to find a synonym.
Context: ...ompt changes. No extra context windows. No new tools to learn. ```bash curl -fsSL...
(ENGLISH_WORD_REPEAT_BEGINNING_RULE)
🤖 Prompt for AI Agents
Verify each finding against the current code and only fix it if needed.
In `@README.md` around lines 3 - 13, The README hero is missing the demo visual
requested in issue `#53`; add a short terminal GIF or screenshot showing
`supermodel watch` output plus the agent reading `.graph` files to demonstrate
the flow — embed it in the hero section right after the rewritten headline (the
block starting at line 3), include alt text and a one-line caption like "Live
demo: supermodel watch + agent using .graph files", and ensure the image file
(GIF/PNG) is added to the repo and referenced by its filename so the README
renders the demo.
|
|
||
| | | | | ||
| |---|---| | ||
| | **Website** | [supermodeltools.com](http://supermodeltools.com) | |
There was a problem hiding this comment.
Use HTTPS for the website link.
Line 198 uses http://supermodeltools.com while the rest of the doc is HTTPS. Prefer HTTPS here for consistency and trust.
Proposed fix
-| **Website** | [supermodeltools.com](http://supermodeltools.com) |
+| **Website** | [supermodeltools.com](https://supermodeltools.com) |📝 Committable suggestion
‼️ IMPORTANT
Carefully review the code before committing. Ensure that it accurately replaces the highlighted code, contains no missing lines, and has no issues with indentation. Thoroughly test & benchmark the code to ensure it meets the requirements.
| | **Website** | [supermodeltools.com](http://supermodeltools.com) | | |
| | **Website** | [supermodeltools.com](https://supermodeltools.com) | |
🤖 Prompt for AI Agents
Verify each finding against the current code and only fix it if needed.
In `@README.md` at line 198, Update the website link in README.md to use HTTPS
instead of HTTP: replace the string "http://supermodeltools.com" with
"https://supermodeltools.com" so the table row "| **Website** |
[supermodeltools.com](http://supermodeltools.com) |" consistently uses HTTPS and
matches the rest of the documentation.
Closes #53 #54
Complete README rewrite focused on outcomes over features.
New headline: "Your AI agent knows your entire codebase — not just the files it's looking at."
New sections:
Restructured:
🤖 Generated with Claude Code
Summary by CodeRabbit
setupandwatchcommands.