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/annihilater/token-stats/agents-mdgit clone --depth 1 https://github.com/Annihilater/token-statsWhat 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 | $0.03374 | $0.03374 |
| Opus 5 | $0.01687 | $0.01687 |
| Sonnet 5 | $0.00675 | $0.00675 |
| Haiku 4.5 | $0.00337 | $0.00337 |
Grade A, and why
token-stats AGENTS.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.
This is a copy
95% identical to tokscale AGENTS.md — 23 lines differ, which has more behind it and is treated as the original. This page carries a canonical link to it rather than competing with it.
How it starts
The opening of the file, as written. The whole thing — 262 lines — stays where its author put it; the contents beside it link to each section on GitHub.
AI Agent Guidelines
This document provides context for AI agents working on the Tokscale project.
Maintaining AGENTS.md Files
When updating AGENTS.md files, follow these principles:
- No hardcoded counts — Don't write "10 crates" or "5 modules"; these become outdated instantly
- No exhaustive lists — Prefer dynamic commands (
ls crates/) over maintaining complete lists - Document constraints, not descriptions — Focus on non-obvious behaviors, gotchas, and cross-crate dependencies
- Use nested AGENTS.md — Place crate-specific details in
crates/{name}/AGENTS.md, not here - Verify before documenting — Grep/read the code to confirm claims are accurate
- Delete outdated info — Outdated docs are worse than no docs
Authoring PR/Issue Content via gh CLI
When writing PR bodies, issue bodies, or comments through gh pr create, gh issue comment, etc.: prefer --body-file with a written-out file over inline --body heredocs.
Why: inline --body "$(cat <<'EOF' ... EOF)" patterns lead to recurring mistakes where backticks are written as \` (incorrectly escaping them inside a single-quoted heredoc where no escaping is needed). The literal \` then renders in GitHub markdown as backslash + backtick instead of inline code formatting.
DO:
- Write the body to
/tmp/pr-body.md(or similar) with theWritetool, thengh pr create --body-file /tmp/pr-body.md - For comment edits:
gh api -X PATCH repos/<owner>/<repo>/issues/comments/<id> -F body=@/tmp/comment.md
DON'T:
- Use
gh pr create --body "$(cat <<'EOF' ... \... ` ... EOF)"— single-quoted heredoc means backticks are already literal; escaping with `` \`` is wrong and renders incorrectly. - Use double-quoted heredoc for bodies containing backticks unless you intend command substitution.
- Hard-wrap paragraphs at ~80 columns inside PR/issue/comment bodies. GitHub's markdown collapses single newlines into spaces so the rendered output looks fine, but the raw markdown view is what reviewers and authors edit in, and mid-sentence line breaks read as if the prose was chopped. Write each paragraph as one continuous line and let the renderer wrap it. Same rule for blockquotes, list items that span multiple lines, and table cells. Hard wraps are still fine inside fenced code blocks where preserving line layout matters.
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 · 262 lines · 3,374 tokens per session scan A f97651f9d0f9
token-stats AGENTS.md is an instructions file published in the GitHub repository Annihilater/token-stats (5 stars, last pushed 1mo ago), licensed MIT. It adds 3,374 tokens to every session, about $0.0169 per session on Opus 5. A static security scan graded it A with 0 findings. It is 95% identical to tokscale AGENTS.md, differing in 23 lines, and is treated as a copy.
Other instructions, from other repositories
tokentelemetry AGENTS.md
AGENTS.md instructions for VasiHemanth/tokentelemetry, covering agents.md, project, common commands and project layout.
toki AGENTS.md
Instructions for choi138/toki, covering workspace rules, source of truth, toki workflow, required checks and global safety rules.
agenttop CLAUDE.md
Instructions for vicarious11/agenttop, covering agenttop — ai usage dashboard & optimizer, project overview, architecture, layers and key directories.
awesome-copilot-id AGENTS.md
Instructions for GulajavaMinistudio/awesome-copilot-id, covering communication, explanation and documentation, markdown formatting, user communication style and workflow & methodology.
hiveshare CLAUDE.md
Instructions for KB-perByte/hiveshare, covering hiveshare — claude.md, repo layout, build & run, key env vars (server) and naming: hive vs memory.
llm-safe-haven CLAUDE.md
Instructions for pleasedodisturb/llm-safe-haven, covering llm safe haven, what this is, project structure, tdd — non-negotiable (adopted 2026-08-17) and the contract.