docs: fix documentation inconsistencies and add missing test flags

[wtfsayo]

Summary

• Change npm install to bun add in packages/server/README.md (project uses bun exclusively)
• Replace broken BUILD_SYSTEM.md link with inline explanation in packages/core/README.md
• Add --port and --skip-type-check flags to ElizaOS test command documentation in CLAUDE.md

Changes

• packages/server/README.md: Fix package manager command
• packages/core/README.md: Fix dead link to non-existent file
• CLAUDE.md: Document missing CLI flags

Test plan

• Changes are documentation-only
• No code changes to verify

🤖 Generated with Claude Code

---

Note

Documentation-only updates across project docs.

• CLI docs: Adds --port and --skip-type-check flags to elizaos test command in CLAUDE.md
• Core README: Replaces dead BUILD_SYSTEM.md reference with a clear note that conditional exports select the correct build automatically
• Server README: Standardizes installation to bun by replacing npm install with bun add for @elizaos/server

^{Written by Cursor Bugbot for commit 856a405. This will update automatically on new commits. Configure here.}

Greptile Summary

• Fixes documentation inconsistencies by correcting package manager commands and broken links across multiple documentation files
• Enhances multi-step execution system with parameter support, retry logic, and improved streaming for more reliable agent workflows
• Adds comprehensive test coverage for multi-step workflow retry mechanisms and parameter extraction functionality

Important Files Changed

Filename	Overview
packages/core/src/services/default-message-service.ts	Major enhancement to multi-step execution with retry logic, parameter handling, streaming improvements, and fallback mechanisms
packages/plugin-bootstrap/src/providers/actions.ts	Adds parameter schema formatting and exposes actionsWithParams for enhanced action context in multi-step workflows
packages/core/src/prompts.ts	Updates multi-step decision template to support parameters and switches to richer action context with actionsWithParams
packages/plugin-bootstrap/src/__tests__/multi-step.test.ts	Adds 183 lines of comprehensive test coverage for retry logic, parameter extraction, and error handling in multi-step workflows

Confidence score: 4/5

• This PR appears safe to merge with some complexity around the multi-step execution changes
• Score reflects the significant architectural improvements to multi-step workflows but includes complex retry logic and parameter handling that requires careful attention
• Pay close attention to the default-message-service.ts file which contains substantial changes to core message processing logic including retry mechanisms and streaming improvements

Sequence Diagram

sequenceDiagram
    participant User
    participant CLI as elizaos CLI
    participant TestRunner as ElizaOS TestRunner
    participant ComponentTests as Component Tests
    participant E2ETests as E2E Tests
    participant Agent as Agent Runtime

    User->>CLI: "elizaos test [path]"
    CLI->>CLI: "Detect project type"
    CLI->>CLI: "Parse command options (--skip-build, --port, etc.)"
    
    alt Component Tests
        User->>CLI: "elizaos test -t component"
        CLI->>ComponentTests: "Run bun test"
        ComponentTests->>ComponentTests: "Test individual modules"
        ComponentTests-->>CLI: "Test results"
    else E2E Tests
        User->>CLI: "elizaos test -t e2e"
        CLI->>TestRunner: "Initialize ElizaOS TestRunner"
        TestRunner->>Agent: "Start agent runtime"
        Agent->>Agent: "Initialize server, database, plugins"
        TestRunner->>TestRunner: "Run integration tests"
        TestRunner-->>CLI: "Test results"
        CLI->>Agent: "Stop agent runtime"
    else All Tests (Default)
        CLI->>ComponentTests: "Run component tests"
        ComponentTests-->>CLI: "Component results"
        CLI->>TestRunner: "Run E2E tests"
        TestRunner->>Agent: "Start agent runtime"
        TestRunner->>TestRunner: "Run integration tests"
        TestRunner-->>CLI: "E2E results"
        CLI->>Agent: "Stop agent runtime"
    end
    
    CLI-->>User: "Test results and exit code"

Loading

Context used:

• Context from dashboard - CLAUDE.md (source)
• Context from dashboard - .cursorrules (source)