Estimate tokens

Skill kyuna0312/context_forge/skills/estimate-tokens

This skill should be used when the user says "estimate token usage", "how many tokens am I using", "token cost of my setup", "audit token budget", "what's eating my context", "token usage breakdown", or "how expensive is my session".From its SKILL.md

Install

npx -y skills add kyuna0312/context_forge --skill estimate-tokens

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

3.1 KB, as published. Nobody here has run it

Estimate Tokens

Audit and report approximate token usage across all context sources: claude.md, skills, memory files, plugins, and conversation history.

Token Estimation Formula

1 word ≈ 1.3 tokens
1 line of code ≈ 10-15 tokens
1KB of text ≈ 200-300 tokens

Audit Procedure

Step 1: Scan claude.md Files

wc -w ~/.claude/CLAUDE.md 2>/dev/null || echo "No global CLAUDE.md"
wc -w ./CLAUDE.md 2>/dev/null || echo "No project CLAUDE.md"

Report: CLAUDE.md: ~[words] words ≈ [words * 1.3] tokens

Step 2: Scan Installed Skills

find ~/.claude/plugins -name "SKILL.md" -exec wc -w {} \; 2>/dev/null | sort -rn | head -20

Skill bodies lazy-load — only frontmatter descriptions are a constant cost (~50–150 tokens each, every session). Report description count × ~100 as the constant estimate; body word counts × 1.3 are the per-invocation cost.

Step 3: Scan Memory Files

wc -w ~/.claude/memory/*.md 2>/dev/null
find .remember -name "*.md" -exec wc -w {} \; 2>/dev/null

Step 4: Count Active Plugins

ls ~/.claude/plugins/ 2>/dev/null | wc -l

Each active plugin = at minimum plugin.json metadata overhead.

Step 5: Generate Report

Output formatted breakdown:

TOKEN USAGE AUDIT
=================
claude.md (global):    ~[N] tokens
claude.md (project):   ~[N] tokens
Skill descriptions:    ~[N] tokens  ([X] skills, constant)
Memory files:          ~[N] tokens
─────────────────────────────────
TOTAL CONSTANT COST:   ~[N] tokens per response

Top token consumers:
1. [skill-name]: ~[N] tokens
2. [skill-name]: ~[N] tokens
3. [memory-file]: ~[N] tokens

Recommendation: [action to reduce]

Optimization Priorities

Based on audit results, recommend in order:

claude.md > 500 words → Run optimize-claudemd skill (usually the biggest constant cost)
Memory files > 1,000 words → Archive old entries
10+ installed plugins → /plugin remove project-irrelevant ones (each skill's description is constant overhead)
Frequently-invoked skill with 3,000+ word body → Move detail to references/ files

LTX Schema

Emit structured output as LTX rows when reporting per-source token estimates.

@v1:source|words|tokens|status

Field	Description
`source`	File or category (e.g. `~/.claude/CLAUDE.md`, `skills`)
`words`	Raw word count
`tokens`	Estimated tokens (`words * 1.3`, rounded)
`status`	`ok`, `warn` (approaching limit), `critical` (over limit)

Example:

@v1:source|words|tokens|status
~/.claude/CLAUDE.md|850|1105|critical
./CLAUDE.md|320|416|ok
skills|2400|3120|warn

Additional Resources

references/token-benchmarks.md — Token cost benchmarks for common setups