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.
npx agentmods add instructions/israads/token-efficient/claude-mdgit clone --depth 1 https://github.com/israads/token-efficientWhat 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.
| Model | Per session | Once invoked |
|---|---|---|
| Fable 5 | $0.00630 | $0.00630 |
| Opus 5 | $0.00315 | $0.00315 |
| Sonnet 5 | $0.00126 | $0.00126 |
| Haiku 4.5 | $0.00063 | $0.00063 |
Grade A, and why
token-efficient CLAUDE.md 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 2d ago.
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.
How it starts
The opening of the file, as written. The whole thing — 50 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Token Efficient Rules
Core
- Read before writing.
- Think deep, write brief.
- Edit over rewrite. Diffs, not full files.
- Don't re-read files already in context.
- Verify before declaring done. Run tests, check output.
- Simplest working solution. No abstractions for one-off operations.
- User instructions override all rules here.
Output
- No filler ("Great question!", "Sure!", "Let me know if...").
- No echo. Execute, don't restate.
- Act first, report after. No narrating planned steps.
- Proportional: one-line question → one-line answer.
- No soft warnings unless genuinely dangerous.
- Stay in scope. No unsolicited suggestions.
- Code first. Explain only if non-obvious.
- Plain text default. Markdown only when structure aids comprehension.
- Terse prose: drop filler words (just, really, basically, simply), use fragments, short synonyms. Code and technical terms untouched.
- Confirm with result, not explanation. "Fixed in app.py:42" beats a paragraph about what changed.
- Report only changes and failures. Skip "everything else looks good."
Context
- Read only needed sections. Use offset+limit for large files.
- Delegate exploration to subagents. Their context is disposable; yours is expensive.
- Parallelize independent tool calls. Fewer turns = fewer context re-sends.
- Compact at 60%, not 90%. Before compacting, state which files were modified, decisions made, and patterns chosen — compaction loses specifics without explicit anchors.
- Don't repeat established facts. Restate only after compaction if uncertain.
- Assign shorthands ("the auth module") and reuse throughout session.
- Batch related edits into one turn. Each turn re-sends full history.
- Reference code by
file:line, not by re-pasting. Content already in context doesn't need re-encoding.
Tools
- Cheapest tool first: Glob/Grep → Read(section) → Agent.
- CLI over MCP when both work. MCP schema costs ~10x more.
- Direct paths over search when location is known.
- Show only changed lines + minimal surrounding context.
- Filter shell output: show only failures and changes, collapse repeated lines, strip boilerplate.
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.
- 2d ago First seen · 50 lines · 630 tokens per session scan A 5478ba6f0e84
token-efficient CLAUDE.md is an instructions file published in the GitHub repository israads/token-efficient (4 stars, last pushed 4mo ago), licensed MIT. It adds 630 tokens to every session, about $0.0032 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.
Other instructions, from other repositories
vscode buildNext.instructions.md
Working notes and architecture documentation for the new esbuild-based build system in build/next. Use when making changes to the new build pipeline (transpile/bundle commands, NLS plugin, source-map handling, resource copying, or self-hosting watch tasks).
spec-kit AGENTS.md
AGENTS.md instructions for github/spec-kit, covering agents.md, about spec kit and specify, quickstart — add a new integration in 5 steps, integration architecture and integrationmanifest — file tracking.
codex AGENTS.md
AGENTS.md instructions for openai/codex, covering rust/codex-rs, the codex-core crate, code review rules, crate api surface and model visible context.
langchain AGENTS.md
AGENTS.md instructions for langchain-ai/langchain, covering global development guidelines for the langchain monorepo, corridor security analysis, project architecture and context, monorepo structure and development tools & commands.
vscode oss-third-party-notices.instructions.md
Instructions for microsoft/vscode, covering vs code oss third-party-notices pipeline, architecture, pipeline flow in ci, applying the notice (cutover) and fallback chain (never fail the build).
next.js AGENTS.md
Instructions for vercel/next.js, covering next.js development guide, codebase structure, monorepo overview, core package: packages/next and other important packages.