⚡ Bolt: Optimize buildContext and cache knowledge retrieval

[SuvenSeo]

💡 What

• Removed the episodic_memory database fetch from the buildContext function.
• Implemented a 5-minute TTL in-memory cache for fetchRelevantKnowledge results, keyed by normalized user message and sorted keywords.
• Cleaned up dead functions (selectConversationLines, compressVerboseContent, scoreEpisodeForContext, isContextNoiseEpisode) and unused constants.

🎯 Why

• The episodic_memory fetch was redundant because conversation history is already managed via the messages array passed to the LLM.
• fetchRelevantKnowledge involves an expensive LLM-based semantic reranking step; caching results for identical queries within a short window significantly improves performance and reduces token usage during tool loops.

📊 Impact

• Reduces 1 database query per AI interaction iteration.
• Eliminates redundant LLM-based reranking calls (saving ~500ms-1s per cached hit).
• Reduces overall system prompt construction time and memory overhead by removing dead processing logic.

🔬 Measurement

• Verified via npm test (32/32 passing).
• Manual trace confirms that episodes was unused in buildContext and that the cache is correctly populated and hit for repeat queries.

---

PR created automatically by Jules for task 16973735803735751205 started by @SuvenSeo

[google-labs-jules]

👋 Jules, reporting for duty! I'm here to lend a hand with this pull request.

When you start a review, I'll add a 👀 emoji to each comment to let you know I've read it. I'll focus on feedback directed at me and will do my best to stay out of conversations between you and other bots or reviewers to keep the noise down.

I'll push a commit with your requested changes shortly after. Please note there might be a delay between these steps, but rest assured I'm on the job!

For more direct control, you can switch me to Reactive Mode. When this mode is on, I will only act on comments where you specifically mention me with @jules. You can find this option in the Pull Request section of your global Jules UI settings. You can always switch back!

New to Jules? Learn more at jules.google/docs.

---

For security, I will only act on instructions from the user who triggered this task.

[vercel]

The latest updates on your projects. Learn more about Vercel for GitHub.

Project	Deployment	Actions	Updated (UTC)
seo-os-agent	Ready	Preview, Comment	May 31, 2026 8:01pm

[Copilot]

Pull request overview

Optimizes buildContext by removing a redundant episodic_memory DB fetch (history is already in the messages array) and adds a 5-minute TTL in-memory cache for fetchRelevantKnowledge to avoid repeated LLM-based semantic reranks within tool-use loops. Also removes now-unused helpers and constants.

Changes:

• Drop episodic_memory query and episodes destructuring from buildContext.
• Cache fetchRelevantKnowledge results by normalized message + sorted keywords for 5 minutes.
• Delete dead helpers (selectConversationLines, compressVerboseContent, scoreEpisodeForContext, isContextNoiseEpisode) and related constants.

Reviewed changes

Copilot reviewed 2 out of 2 changed files in this pull request and generated no comments.

File	Description
frontend/src/lib/services/context.js	Removes episodic memory fetch, caches knowledge retrieval, prunes dead code
.jules/bolt.md	New learning log documenting the optimization rationale

---

💡 Add Copilot custom instructions for smarter, more guided reviews. Learn how to get started.