Merge main and resolve conflicts - integrate CLI auto-shutdown and refactor helpers

[Copilot]

Main branch introduced resolveCliAutoShutdown() and refactored orchestration logic while this PR restructured README and added MCP workflows. Merge conflicts arose in 5 files.

Conflict Resolution Strategy

README.md: Kept PR's restructured content, inserted main's CODEX_CLI_AUTO_SHUTDOWN documentation after build commands section

Binary artifacts (.codex-synaptic/memory.db): Retained PR version via git checkout --ours - SQLite databases auto-rebuild on next run

Documentation (docs/runbooks/autoscaler-daemon-coordination.md): Kept PR's more comprehensive version

Dependencies (package-lock.json): Regenerated using npm install --ignore-scripts to merge both branches' updates

CLI implementation (src/cli/index.ts): Merged main's improvements into PR's refactored structure:

// Integrated main's auto-shutdown resolution
const autoShutdown = options.autoShutdown ?? cliAutoShutdown;

// Applied consistent helper usage
renderBackgroundDaemonStatus(status);  // vs inline status display

// Refactored orchestration into helpers
await executeClassicOrchestration(system, prompt, originalPrompt, config, options, codexContext, codexEnvelope);
await executeGoapStrategy(system, originalPrompt, options);

Import fixes: Added missing GoapExecutor import, removed duplicate setupWorkflowEventHandlers (already exported from hive-mind-helpers), fixed cleanup call pattern to use .cleanup() method

TypeScript compilation verified successful post-merge.

---

💡 You can make Copilot smarter by setting up custom instructions, customizing its development environment and configuring Model Context Protocol (MCP) servers. Learn more Copilot coding agent tips in the docs.

[chatgpt-codex-connector]

You have reached your Codex usage limits for code reviews. You can see your limits in the Codex usage dashboard.

[github-actions]

Codex PR Review

Findings

• P0: codex-synaptic system start now shuts the orchestrator back down immediately. resolveCliAutoShutdown() defaults to true whenever CODEX_CLI_AUTO_SHUTDOWN is unset (src/cli/index.ts:206-215), useSystem honors that default and auto-shuts any newly created session (src/cli/index.ts:787-818), and system start invokes useSystem without overriding autoShutdown (src/cli/index.ts:2753-2759). As a result, running codex-synaptic system start with default env brings the system up and then tears it down before the command exits; the very next system status reports “System not started,” making the command unusable.
• P2: codex-synaptic background status lost critical diagnostics. It now only calls renderBackgroundDaemonStatus (src/cli/index.ts:3034-3037), whose new implementation omits the runtime snapshot lookup and no longer prints CWD, socket/log paths, or readiness (src/cli/index.ts:932-949). On main, the command fetched queryBackgroundRuntimeSnapshot() and surfaced those fields, which operators rely on for daemon debugging.

Suggested Fixes

• Update lifecycle handling so system start keeps the session alive—e.g., either revert the default to non-autoshutdown or pass { autoShutdown: false } when invoking useSystem for commands that are supposed to keep the orchestrator running.
• Restore the background status diagnostics by querying queryBackgroundRuntimeSnapshot() and printing the previously available metadata (cwd, socket, runtime/log paths, readiness) alongside the basic fields.

MCP Usage

• deepwiki – unavailable (missing DEEPWIKI_API_KEY).
• context7 – unavailable (resources/list method not exposed).
• brave – unavailable (handshake failed/connection closed).
• jina – unavailable (missing JINA_API_KEY).
• firecrawl – unavailable (handshake failed/connection closed).