Check claudemd size

Skill kyuna0312/context_forge/skills/check-claudemd-size

This skill should be used when the user says "check claude.md size", "is my claude.md too big", "claude.md token cost", "how big is claude.md", or when triggered automatically on session start to warn about oversized claude.md files.From its SKILL.md

Install

npx -y skills add kyuna0312/context_forge --skill check-claudemd-size

Assembled from the repository path, not quoted from the project. Check it against their README if it does not work.

One thing to look at

3 stars3 stars. Stars are a popularity signal and not a quality one, but at this level it is likely that nobody has read this closely except its author, and you would be relying on your own review.

SKILL.md

2.6 KB, as published. Nobody here has run it

Check Claude.md Size

Measure and report claude.md file sizes. Warn when files exceed optimal token budget. Suggest optimization when needed.

Size Thresholds

Size	Status	Action
< 300 words	Optimal	No action needed
300-600 words	Acceptable	Monitor growth
600-1,000 words	Bloated	Run optimize-claudemd
> 1,000 words	Critical	Immediate optimization needed

Check Procedure

Scan All claude.md Files

# Global
wc -w ~/.claude/CLAUDE.md 2>/dev/null
wc -w ~/.claude/claude.md 2>/dev/null

# Project
wc -w ./CLAUDE.md 2>/dev/null
wc -w ./claude.md 2>/dev/null

Report Format

CLAUDE.MD SIZE CHECK
====================
Global (~/.claude/CLAUDE.md):   [N] words ≈ [N*1.3] tokens  [STATUS]
Project (./CLAUDE.md):          [N] words ≈ [N*1.3] tokens  [STATUS]

Combined constant cost: ~[N] tokens per response

[WARNING if any file is Bloated/Critical]
Run: /context_forge:optimize-claudemd to reduce size

Automatic Session-Start Check

The SessionStart hook (hooks/scripts/session-start.sh) performs this same check automatically — it does not invoke this skill. Use the skill for on-demand checks mid-session; don't re-run it right after session start, the hook already reported. On session start the hook:

Check all claude.md file sizes
If any file > 600 words: print warning
If any file > 1,000 words: print critical warning with optimization command

Warning format:

⚠ TOKEN SAVER: claude.md is [N] words ([STATUS])
  Run /context_forge:optimize-claudemd to reduce by ~[%]%

What Makes claude.md Grow

Common causes of bloat:

Accumulated instructions over time without pruning
Paste of full API docs or schemas
Duplicate rules from different sessions
Long prose explanations instead of bullet rules
Examples that should be in references/ files

LTX Schema

Emit structured output as LTX rows when reporting file sizes.

@v1:file|words|tokens|level

Field	Description
`file`	Path to the claude.md file checked
`words`	Raw word count
`tokens`	Estimated tokens (`words * 1.3`, rounded)
`level`	`ok`, `warn` (600–999 words), `critical` (≥1000 words)

Example:

@v1:file|words|tokens|level
~/.claude/CLAUDE.md|850|1105|critical
./CLAUDE.md|320|416|ok