feat: massive overhaul to manage costs (#103)

### Explanation

_No response_

### Issue

Closes #102

### Attestations

- [ ] I have read and agree to the [Code of Conduct](https://docs.nhcarrigan.com/community/coc/)
- [ ] I have read and agree to the [Community Guidelines](https://docs.nhcarrigan.com/community/guide/).
- [ ] My contribution complies with the [Contributor Covenant](https://docs.nhcarrigan.com/dev/covenant/).

### Dependencies

- [ ] I have pinned the dependencies to a specific patch version.

### Style

- [ ] I have run the linter and resolved any errors.
- [ ] My pull request uses an appropriate title, matching the conventional commit standards.
- [ ] My scope of feat/fix/chore/etc. correctly matches the nature of changes in my pull request.

### Tests

- [ ] My contribution adds new code, and I have added tests to cover it.
- [ ] My contribution modifies existing code, and I have updated the tests to reflect these changes.
- [ ] All new and existing tests pass locally with my changes.
- [ ] Code coverage remains at or above the configured threshold.

### Documentation

_No response_

### Versioning

_No response_

Reviewed-on: #103
Co-authored-by: Naomi Carrigan <commits@nhcarrigan.com>
Co-committed-by: Naomi Carrigan <commits@nhcarrigan.com>

src/lib/utils/conversationUtils.ts

@@ -0,0 +1,110 @@
+import type { ConversationSummary } from "$lib/stores/conversations";
+
+/**
+ * Sanitises a string for safe JSON serialization through Tauri IPC.
+ * Removes control characters and lone surrogates that could cause issues
+ * during JSON serialization/deserialization.
+ */
+export function sanitizeForJson(text: string): string {
+  // Remove control characters except for common whitespace (tab, newline, carriage return)
+  // These can cause JSON parsing issues and are rarely meaningful in conversation summaries.
+  // eslint-disable-next-line no-control-regex -- regex uses control character codes
+  let sanitized = text.replace(/[\x00-\x08\x0B\x0C\x0E-\x1F\x7F]/g, "");
+
+  // Remove extended ASCII control chars (C1 control codes)
+  sanitized = sanitized.replace(/[\x80-\x9F]/g, "");
+
+  // Remove lone surrogates (U+D800 to U+DFFF) which cause "unexpected end of hex escape"
+  // errors in serde_json when they appear without proper pairing.
+  // These are invalid in JSON and can cause parse failures.
+  sanitized = sanitized.replace(/[\uD800-\uDFFF]/g, "");
+
+  return sanitized;
+}
+
+/**
+ * Generates a prompt to ask Claude to summarise a conversation.
+ * This can be sent as a user message to get a summary.
+ */
+export function generateSummaryPrompt(conversationContent: string): string {
+  return `Please provide a concise summary of our conversation so far. Focus on:
+1. Key topics discussed
+2. Important decisions or conclusions made
+3. Any ongoing tasks or context that would be helpful to remember
+4. Code changes or files that were modified
+
+Keep the summary brief but comprehensive enough to continue our work in a new session.
+
+Here is our conversation:
+
+${conversationContent}
+
+Please provide the summary now:`;
+}
+
+/**
+ * Generates a context injection message to prepend to a new conversation.
+ * This provides Claude with context from a previous session.
+ */
+export function generateContextInjection(summary: ConversationSummary): string {
+  return `[Previous Session Context]
+The following is a summary from our previous conversation (${summary.messageCount} messages, approximately ${summary.tokenEstimate.toLocaleString()} tokens):
+
+${summary.content}
+
+[End of Previous Context]
+
+Please continue from where we left off, or let me know if you need any clarification about the previous context.`;
+}
+
+/**
+ * Estimates the token count for a given string.
+ * Uses a rough approximation of ~4 characters per token.
+ */
+export function estimateTokens(text: string): number {
+  return Math.ceil(text.length / 4);
+}
+
+/**
+ * Creates a ConversationSummary object from summary content.
+ */
+export function createSummary(
+  content: string,
+  messageCount: number,
+  originalTokenEstimate: number
+): ConversationSummary {
+  return {
+    generatedAt: new Date(),
+    content,
+    messageCount,
+    tokenEstimate: originalTokenEstimate,
+  };
+}
+
+/**
+ * Determines if a conversation should be compacted based on token usage.
+ * Returns true if the conversation is using more than the threshold percentage
+ * of the context window.
+ */
+export function shouldSuggestCompaction(
+  contextTokensUsed: number,
+  contextWindowLimit: number,
+  thresholdPercent: number = 60
+): boolean {
+  if (contextWindowLimit === 0) return false;
+  const utilisationPercent = (contextTokensUsed / contextWindowLimit) * 100;
+  return utilisationPercent >= thresholdPercent;
+}
+
+/**
+ * Formats a token count for display.
+ */
+export function formatTokenCount(tokens: number): string {
+  if (tokens >= 1000000) {
+    return `${(tokens / 1000000).toFixed(1)}M`;
+  }
+  if (tokens >= 1000) {
+    return `${(tokens / 1000).toFixed(1)}K`;
+  }
+  return tokens.toString();
+}