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/ekalb81/agent-odometer/agents-mdgit clone --depth 1 https://github.com/ekalb81/agent-odometerWhat 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 | $0.02866 | $0.02866 |
| Opus 5 | $0.01433 | $0.01433 |
| Sonnet 5 | $0.00573 | $0.00573 |
| Haiku 4.5 | $0.00287 | $0.00287 |
Grade A, and why
agent-odometer 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.
How it starts
The opening of the file, as written. The whole thing — 107 lines — stays where its author put it; the contents beside it link to each section on GitHub.
AGENTS.md
These instructions apply to the entire repository.
Project purpose
Odometer is a local Tauri companion to agent CLI harnesses: the ChatGPT desktop app's Codex experience and Claude Code. It reads each harness's session JSONL files and presents searchable task, turn, token, subagent, and estimated cost data in per-harness tabs. Rust owns filesystem access, parsing, persistence, and Tauri IPC; Svelte owns filtering, presentation, and credit calculations.
Start with README.md for commands and docs/ARCHITECTURE.md for the data flow, contracts, invariants, and known limitations.
Before changing code
- Run
git status --shortand preserve unrelated work. Do not normalize or rewrite existing files as a side effect. - Treat
src-tauri/gen/schemas/as generated. Do not hand-edit it; commit generated changes only when an intentional Tauri capability/configuration change requires them. - Both lockfiles are authoritative and should stay committed. Use
npm ciand Cargo's--lockedflag for reproducible validation. - Never commit real Codex sessions, prompts, local paths, credentials, or platform config. Parser fixtures must be synthetic or thoroughly redacted.
Architecture boundaries
- Keep filesystem access and JSONL parsing in Rust. The frontend should receive typed data through Tauri commands/events, not read session files directly.
- Put Rust commands in
src-tauri/src/commands.rs, register them insrc-tauri/src/lib.rs, and expose typed frontend wrappers insrc/lib/ipc.ts. - Keep Rust serialized structs and
src/lib/types.tssynchronized. Add backward-compatible Serde defaults when persisted or historical data may omit a new field. - Event names are contracts:
session-updated(payload:SessionSummary),session-removed,scan-progress(payload:ScanStatus),history-progress(payload:HistoryStatus),config-updated, andrates-updated. Update every producer and listener together. Full sessions travel only throughget_session_details; keepSessionSummaryfree ofturns/tokens_history— the split exists because full sessions measured ~200 MB across a real corpus. history-progress(#116) mirrorsscan-progress's contract for the durable-history archive:AppStatestarts in a thirdPendingstate (distinct from an archive that is genuinelyUnavailable) untilcommands::spawn_history_openresolves it on a background thread, emitting one event per migration step plus a terminalready/unavailableevent. The frontend callsget_history_statusonce on mount (events may have already fired) and then follows the event, matchingscan-progress/get_scan_status.commands::sessions_in_ranges— the ledger accounting authority — checksAppState::history_readiness()directly and returns an explicit "still preparing" error whilePending, rather than falling back to a partial in-memory answer that would look complete.- Use the established Svelte 5 rune style (
$state,$derived,$effect). Module-level rune state belongs in*.svelte.tsfiles. - Keep Tauri capabilities minimal. Do not add remote content, network access, shell execution, or broader capabilities without an explicit requirement and a security review. Current exceptions:
updater:defaultandprocess:allow-restartexist solely for the in-app auto-updater.
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 · 107 lines · 2,866 tokens per session scan A 685232fbd47e
agent-odometer AGENTS.md is an instructions file published in the GitHub repository ekalb81/agent-odometer (2 stars, last pushed 5d ago), licensed MIT. It adds 2,866 tokens to every session, about $0.0143 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
codex-usage-desktop AGENTS.md
Instructions for itvincent-git/codex-usage-desktop, covering agents.md, 1. think before coding, 2. simplicity first, 3. surgical changes and 4. goal-driven execution.
Token-Dashboard CLAUDE.md
Instructions for Arylmera/Token-Dashboard, covering claude.md, project overview, status, architecture and data source.
agent-usage-desktop AGENTS.md
Instructions for hongshuo-wang/agent-usage-desktop, covering agents.md, build & run, desktop app (tauri), macos arm64 example and macos x8664.
codex-usage-desktop GEMINI.md
Instructions for itvincent-git/codex-usage-desktop, a project described as: A local-first desktop dashboard for macOS and Windows that turns Codex CLI and app logs into clear insights on token usage, estimated costs, models, and projects.
agent-usage-desktop CLAUDE.md
Instructions for hongshuo-wang/agent-usage-desktop, a project described as: A desktop app for local usage, cost, throughput, and session analytics across Claude Code, Codex, OpenCode, and OpenClaw. | 在本地统一统计和分析 Claude Code、Codex、OpenCode 与 OpenClaw 的用量、费用、吞吐和会话。.
nextclaw AGENTS.md
Instructions for Peiiii/nextclaw, covering nextclaw ai 常驻内核, 产品愿景, 沟通与推进, 深思与目标模式 and 协作与 git 安全.