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/marve10s/better-fullstack/claude-mdgit clone --depth 1 https://github.com/Marve10s/Better-FullstackWhat 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 | $0.01974 | $0.01974 |
| Opus 5 | $0.00987 | $0.00987 |
| Sonnet 5 | $0.00395 | $0.00395 |
| Haiku 4.5 | $0.00197 | $0.00197 |
Grade B, and why
Better-Fullstack CLAUDE.md scanned grade B with 1 finding 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.
Reads agent configuration directoriesmediumAgent snooping
.claude/, .codex/, .gemini/ hold keys, settings and other credentials a mod has no legitimate need for.
- Mechanics: GPT models are only reachable through the Codex CLI - `codex exec` / `codex review` (my ~/.codex/config.toml defaults to gpt-5.6-sol at effort high). Pick a tier per call with `codex exec -m gpt-5.6-luna|ter How it starts
The opening of the file, as written. The whole thing — 67 lines — stays where its author put it; the contents beside it link to each section on GitHub.
CLAUDE.md
Canonical agent instructions live in AGENTS.md. This file exists so Claude Code loads those instructions automatically.
Read AGENTS.md for all project conventions, guidelines, and workflow rules.
Code Style
- Always strive for concise, simple solutions.
- If a problem can be solved in a simpler way, propose it.
General preferences
- If asked to do too much work at once, stop and state that clearly.
- If computer use is helpful for completing or verifying work, shell out to gpt-5.6 with Codex for it.
Picking the right models for workflows and subagents
Rankings, higher = better. Cost reflects what I actually pay (OpenAI is near-free for me due to a deal, and Claude models are close behind), not list price. Speed is throughput and time-to-answer. Intelligence is how hard a problem you can hand the model unsupervised. Taste covers UI/UX, code quality, API design, and copy.
| model | cost | speed | intelligence | taste |
|---|---|---|---|---|
| gpt-5.6-sol | 9 | 4 | 9 | 6 |
| gpt-5.6-terra | 9 | 6 | 8 | 5 |
| gpt-5.6-luna | 9 | 9 | 6 | 5 |
| sonnet-5 | 5 | 7 | 7 | 7 |
| opus-5 | 8 | 5 | 8 | 8 |
| fable-5 | 2 | 2 | 9 | 9 |
GPT-5.6 tier notes (GPT-5.5 is retired - don't use it):
- Sol - the GPT flagship. Frontier-level agentic coding; writes tight, restrained code. Use for hard unsupervised problems where taste doesn't matter. Supports
maxandultrareasoning efforts (ultraspawns subagents - always pair withrollout_token_budget). User-facing work stays on Claude models. - Terra - the everyday workhorse. Use for routine well-specified implementation and mid-weight investigation where Sol is overkill. Since my OpenAI cost is flat, its niche is mostly rate-limit headroom.
- Luna - smallest and fastest tier; weakest on open-ended reasoning. Default for bulk mechanical work: migrations, sweeps across many files, data grinding, first-pass verification - anywhere throughput and latency matter more than depth.
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.
- 3d ago First seen · 67 lines · 1,974 tokens per session scan B 782f86a008af
Better-Fullstack CLAUDE.md is an instructions file published in the GitHub repository Marve10s/Better-Fullstack (726 stars, last pushed 3d ago), licensed MIT. It adds 1,974 tokens to every session, about $0.0099 per session on Opus 5. A static security scan graded it B with 1 finding (reads agent configuration directories). No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-30.
Other instructions, from other repositories
drizzle-cube CLAUDE.md
Instructions for cliftonc/drizzle-cube, covering claude.md, project structure, agent documentation context, database support and essential commands.
flarestarter CLAUDE.md
Instructions for FlareStarter/flarestarter: This file provides guidance to Claude Code (claude.ai/code) when working with code in this repository.
better-drizzle GEMINI.md
Instructions for almeidazs/better-drizzle: Use AGENTS.md as the repository-wide source of truth.
better-drizzle CLAUDE.md
Instructions for almeidazs/better-drizzle: Use AGENTS.md as the repository-wide source of truth.
evjs AGENTS.md
Instructions for afx-team/evjs, covering agents.md, start here, source map, working rules and validation.
ai-coding-primer CLAUDE.md
Instructions for ammonhaggerty/ai-coding-primer, covering claude.md, repository structure, guidebook chapters, constraints and writing tone.