fix: strip all accumulated tags in stripTagPrefix to prevent tag accumulation

[tomolom]

Summary

Fixes tag accumulation bug where stripTagPrefix only removed a single tag, causing §N§ symbols to accumulate when content is processed multiple times during transform passes, compartment compaction, or message replay.

Changes

Changed TAG_PREFIX_REGEX in tag-content-primitives.ts:

// Before: Only matches ONE tag
const TAG_PREFIX_REGEX = /^§\d+§\s*/;

// After: Matches one or more consecutive tags  
const TAG_PREFIX_REGEX = /^(?:§\d+§\s*)+/;

Problem

The original regex only stripped the outermost tag prefix, leaving previously accumulated tags intact. Each transform pass would prepend a new tag on top of old ones, resulting in unreadable messages like:

§4687§ §4686§ §4685§ ... actual content

Solution

The new regex uses a non-capturing group with + quantifier to match all consecutive tag prefixes at the start of the string. This ensures stripTagPrefix removes all accumulated cruft before prependTag adds a single fresh tag.

Testing

This change maintains backward compatibility - existing single-tag content works identically, and multi-tag content is now properly normalized to a single tag.

Fixes #11

[Copilot]

Pull request overview

Fixes a tag prefix normalization bug in the magic-context tagging utilities so repeated transform passes don’t accumulate multiple §<id>§ prefixes on the same content.

Changes:

• Update the tag-prefix stripping regex to remove all consecutive §<id>§ prefixes at the start of a string (not just the first).
• Ensure prependTag() consistently results in a single tag prefix even when content has already been tagged multiple times.

---

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