lean-stats

A command that measures output tokens in Claude Code session transcripts and compares two runs using the same prompts.

In plain words
What is it for?
It reports tokens and the estimated compressible prose limit for one transcript, or the measured difference between lean-off and lean-on runs.
Why use it?
It replaces guesses about shorter answers with measurements from the actual session files.

Command

Install

Getting it into your agent

One page per mod, every tool's command on it. A separate URL per tool would split the same page into five that compete with each other.

agentmods
npx agentmods add commands/ousid/lean/lean-stats
Clone the repo
git clone --depth 1 https://github.com/ousid/lean
Per session 9 Only the description is in the session, so the agent can decide to use it. The body loads when it is invoked.
When invoked 196 The whole file, excluding the scripts and references it only reads on demand.
Security scan A 0 findings. Scan, not verified.
Origin original No closer match found in the catalogue.
Token cost

What it costs to keep this loaded

Counted locally with the o200k_base tokenizer, which is exact for GPT models; Claude uses its own tokenizer and its counts differ. Treat this as one consistent yardstick across the catalogue rather than a bill. Prices are per million input tokens.

ModelPer sessionOnce invoked
Fable 5 $0.00009 $0.00196
Opus 5 $0.00005 $0.00098
Sonnet 5 $0.00002 $0.00039
Haiku 4.5 $0.00001 $0.00020

Measured yesterday against content hash d9e64b3b6db6, method: parsed. Prices are Anthropic first-party input rates as of 2026-08-30, from the pricing page.

Security

Grade A, and why

lean-stats scanned grade A with 0 findings against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured yesterday.

A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.

Nothing flagged

None of the 26 patterns this scan looks for appear in this file: no shell pipes, no recursive deletes, no credential paths, no hidden text, no instruction-override or anti-refusal phrasing, no agent-config snooping. That is not a guarantee, it is the absence of the things that are checkable.

commands/lean-stats.md · 14 lines

What it actually says

Run the lean-stats tool and show me the output. Do not invent numbers.

Steps:

  1. Find my most recent Claude Code transcripts if I did not pass paths: ls -t ~/.claude/projects/*/*.jsonl | head.
  2. Run node "${CLAUDE_PLUGIN_ROOT}/bin/lean-stats.mjs" $ARGUMENTS.
    • report <session.jsonl> gives real output tokens plus the compressible prose ceiling for one session.
    • diff <baseline.jsonl> <lean.jsonl> gives the measured saving between a lean off run and a lean on run of the SAME prompts.
  3. Print the raw tool output. If it says a value is approx or unmeasured, keep that label. Never round an estimate into a headline number.
Changes

What this file has done since we first saw it

Hashed on every crawl. A supply-chain change to an agent config is a question of when, not whether, so the history is kept rather than the latest state alone.

  1. yesterday First seen · 14 lines · 9 tokens per session scan A d9e64b3b6db6

Subscribe to this mod's changes

lean-stats is a command published in the GitHub repository ousid/lean (4 stars, last pushed 1mo ago), licensed MIT. It adds 9 tokens to every session and 196 once invoked, about $0.0000 per session on Opus 5. A static security scan graded it A with 0 findings. No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-31.