Revamp README for launch - #6
Conversation
Rewrite the CLI README from a reference doc into a launch-ready landing page. Replace the 30-row command table and internal architecture section with a value-prop hero, quick start, templates, how-it-works diagram, and community links. Add the Castari logo SVG to assets/. Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
There was a problem hiding this comment.
💡 Codex Review
Here are some automated review suggestions for this pull request.
Reviewed commit: bcd87e83c9
ℹ️ About Codex in GitHub
Your team has set up Codex to review pull requests in this repo. Reviews are triggered when you
- Open a pull request for review
- Mark a draft as ready
- Comment "@codex review".
If Codex has suggestions, it will comment; otherwise it will react with 👍.
Codex can also answer questions or update the PR. Try commenting "@codex address that feedback".
|
|
||
| | Template | What it does | | ||
| |----------|-------------| | ||
| | `cast init --template default` | General-purpose coding assistant with file and bash tools | |
There was a problem hiding this comment.
Include project name in template init examples
The template quickstart rows use commands like cast init --template default, but init requires a positional project name, so copy/pasting these examples fails with Project name is required instead of scaffolding an agent. The CLI enforces this in initCommand with .argument('[name]') and an early if (!name) ... process.exit(1) check (packages/cli/src/commands/init.ts, lines 246-263), so the README should document cast init <name> --template <template>.
Useful? React with 👍 / 👎.
Replace @castari/sdk section with a "Build Your Agent" section showing how to build with the Claude Agent SDK and deploy with Castari. Remove E2B and Daytona mentions from intro, diagram, and footer. Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
CLAUDE.md is a convention in the starter templates, not a Castari requirement. The only required file is castari.json. Updated all references to accurately describe the agent structure. Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
Replace raw Anthropic SDK example (client.messages.create) with the real Claude Agent SDK (query from @anthropic-ai/claude-agent-sdk). Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
platform.claude.com/docs/en/agent-sdk/overview, not the old docs.anthropic.com path. Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
|
Related: #7 adds a Claude Code plugin and |
Add a callout after the Install section pointing users to the /castari-deploy skill as a faster alternative to manual setup. Related: #7 Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
Lead with the two core differentiators: - Agents get their own computer (real filesystem, shell, packages) - Isolated sandboxes eliminate risk to local machines Reinforced in the "How It Works" steps. Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
Shift from developer/coding-agent framing to the primary use case: building agents that serve end users. Native filesystem enables agents to generate reports, process uploads, and deliver real files. Isolated sandboxes mean each customer session is secure and independent. Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
pnpm/action-setup@v4 now reads the version from package.json's packageManager field automatically. Having both causes a build error. Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
Summary
assets/for self-contained rendering on GitHubTest plan
🤖 Generated with Claude Code