Engram v1.7.1 — UTF-8 Truncation Fix

Fixes a bug where truncating long observations could produce invalid UTF-8 by cutting multi-byte characters mid-rune.

Bug Fix

• Safe truncation — truncate() now counts runes instead of bytes, ensuring observation content is always valid UTF-8 after trimming (#26)

Docs

• Added Memory Protocol Full Text section to DOCS.md (#30)
• Updated VS Code user-profile MCP and instruction paths