Skip to content

Escape LinkedIn reserved characters so posts aren't truncated - #38

Merged
PriyeshDave merged 1 commit into
mainfrom
fix/linkedin-reserved-chars
Nov 9, 2026
Merged

PriyeshDave merged 1 commit into
mainfrom
fix/linkedin-reserved-chars

Conversation

@PriyeshDave

Copy link
Copy Markdown
Owner

Problem

The LinkedIn post for Agentic memory patterns in the wild was cut off mid-bullet ("Vector store memory is fast"). LinkedIn's Posts API parses commentary as little-text, where ( ) [ ] { } < > @ # * _ ~ | \ are reserved syntax; an unescaped one silently truncates everything after it. linkedin_copy.py passed LLM output through untouched, so any bullet with e.g. "(10–40ms)" broke the post.

Fix

  • Map reserved characters to fullwidth look-alikes (( )[ ]< > @ …) in the headline, context lines, bullets and the attribution line.
  • Hashtags keep a literal # (so they auto-link) and are reduced to alphanumerics.

Tested locally with a stubbed LLM response containing every reserved character: none remain in the body, hashtags intact.


Generated by Claude Code

LinkedIn parses commentary as little-text; reserved ASCII like ( ) [ ] < > @ silently
truncates the post. Swap them for fullwidth look-alikes in headline, context and
bullets; sanitize hashtags to alphanumerics so they still auto-link.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01Wd36GnTiTuBTpbteZ4P5QC
@PriyeshDave
PriyeshDave merged commit 4818843 into main Nov 9, 2026
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

2 participants