AIUsageTracker AGENTS.md

AIUsageTracker AGENTS.md is an instructions file for Codex, OpenCode from rygel/AIUsageTracker. It costs 8,825 tokens per session, scanned A, original, MIT.

A project instruction file for AI assistants working on a .NET 8 WPF application. It records rules for preserving user settings, handling preference files, testing compatibility, and controlling releases.

In plain words
What is it for?
Guiding work on preference deserialization, compatibility between string and numeric settings, required tests, data-preserving migrations, and release permissions.
Why use it?
It reduces the risk of migrations or code changes deleting preferences, mishandling older settings, or creating releases without approval.

Instructions file for CodexOpenCode

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 instructions/rygel/aiusagetracker/agents-md
Clone the repo
git clone --depth 1 https://github.com/rygel/AIUsageTracker

Made for: Codex, OpenCode.

Wrote this? Show the measurements

A badge with what this costs and how it scanned, read live from this page, so it follows the numbers instead of freezing them. Markdown for a README, HTML for a documentation site or a project page.

agentmods badge for AIUsageTracker AGENTS.md

README.md
[![agentmods](https://agentmods.dev/badge/instructions/rygel/aiusagetracker/agents-md.svg)](https://agentmods.dev/instructions/rygel/aiusagetracker/agents-md)
Your own site
<a href="https://agentmods.dev/instructions/rygel/aiusagetracker/agents-md"><img src="https://agentmods.dev/badge/instructions/rygel/aiusagetracker/agents-md.svg" alt="Measured on agentmods" height="20"></a>
Per session 8,825 This file is loaded in full into every session.
When invoked 8,825 The same file — it is already loaded in full.
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 $0.08825 $0.08825
Opus 5 $0.04412 $0.04412
Sonnet 5 $0.01765 $0.01765
Haiku 4.5 $0.00882 $0.00882

Measured 3d ago against content hash 076d23166937, method: parsed. Prices are Anthropic first-party input rates as of 2026-08-30, from the pricing page.

Security

Grade A, and why

AIUsageTracker 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 3d 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.

AGENTS.md · 803 lines

How it starts

The opening of the file, as written. The whole thing — 803 lines — stays where its author put it; the contents beside it link to each section on GitHub.

AI Usage Tracker - Monitor Guidelines

This document provides essential information for agentic coding assistants working on this .NET 8.0 WPF application.

Critical Rules

NEVER Wipe User Settings (CRITICAL)

  • User preferences are sacred. NEVER allow a deserialization failure to reset ALL preferences to defaults.
  • AppPreferences.Deserialize MUST catch JsonException and retry with lenient options before falling back to defaults. A single incompatible property must NEVER destroy every setting the user has configured.
  • Every enum property on AppPreferences MUST have [JsonConverter(typeof(JsonStringEnumConverter<T>))] so it accepts both string and numeric values on read. This is mandatory for forward/backward compatibility across versions.
  • Mandatory tests — these MUST exist and MUST pass or the build fails:
    • preferences.json with string enum values (e.g. "UpdateChannel": "Beta") must deserialize correctly even if the current code writes numeric.
    • preferences.json with numeric enum values (e.g. "UpdateChannel": 1) must deserialize correctly even if the current code writes strings.
    • A preferences.json with ONE corrupt property must still salvage ALL other valid properties — the corrupt one is skipped, not the entire file.
  • Root cause context: Merge commit 8eb7c163 silently dropped the JsonStringEnumConverter attribute from UpdateChannel. Users who had "UpdateChannel": "Beta" (string) in their preferences.json hit a JsonException on startup, and PreferencesStore.LoadAsync() returned new AppPreferences() — wiping EVERY setting (theme, fonts, thresholds, window position, notification config, hidden providers, everything). This must NEVER happen again.

NEVER Cause Data Loss in Migrations (CRITICAL)

  • Customer data is sacred. A database migration MUST NEVER lose data. Every column must survive every migration step.
  • SQLite table-recreation migrations (CREATE new → INSERT INTO new → DROP old → RENAME) are the #1 data-loss risk. The INSERT SELECT MUST include EVERY column that exists in the source table. No exceptions.
  • Mandatory: When adding a column to provider_history (or any table that ConvertTimestampsToEpochIfNeeded recreates), the column MUST be added to BOTH:
    1. The CREATE TABLE statement in the conversion
    2. The INSERT INTO ... SELECT statement in the conversion
  • Mandatory test: Every table-recreation migration MUST have a test that inserts data WITH the relevant columns populated, runs the migration, and verifies the data survived. A test that only checks column existence is NOT sufficient — it must verify DATA preservation.
  • Single source of truth: The EnsureColumn calls and the conversion's column list must stay in sync. If they drift, data is silently destroyed. The conversion's CREATE TABLE and INSERT SELECT must match the EnsureColumn list exactly.
  • Root cause context: PR #654 added 6 card columns (card_type, window_kind, card_id, group_id, model_name, name) to provider_history. ConvertTimestampsToEpochIfNeeded recreated the table with a hardcoded 15-column INSERT that didn't include them. The columns were added back empty via EnsureColumn AFTER the recreation — permanently wiping card classification data from all historical rows. This caused OpenAI dual bars to disappear.

Read the full file on GitHub · 803 lines

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. 3d ago First seen · 803 lines · 8,825 tokens per session scan A 076d23166937

Subscribe to this mod's changes

AIUsageTracker AGENTS.md is an instructions file published in the GitHub repository rygel/AIUsageTracker (44 stars, last pushed 8d ago), licensed MIT. It adds 8,825 tokens to every session, about $0.0441 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 2026-08-30.