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/eight-acres-lab/openmelon/claude-mdgit clone --depth 1 https://github.com/eight-acres-lab/openmelonWrote 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.
[](https://agentmods.dev/instructions/eight-acres-lab/openmelon/claude-md)<a href="https://agentmods.dev/instructions/eight-acres-lab/openmelon/claude-md"><img src="https://agentmods.dev/badge/instructions/eight-acres-lab/openmelon/claude-md.svg" alt="Measured on agentmods" height="20"></a>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.
| Model | Per session | Once invoked |
|---|---|---|
| Fable 5 | $0.02113 | $0.02113 |
| Opus 5 | $0.01056 | $0.01056 |
| Sonnet 5 | $0.00423 | $0.00423 |
| Haiku 4.5 | $0.00211 | $0.00211 |
Grade A, and why
openmelon 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 5d 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 — 136 lines — stays where its author put it; the contents beside it link to each section on GitHub.
CLAUDE.md — openmelon
Guidance for Claude Code (claude.ai/code) when working in this repo.
What this is
openmelon is a content-creation agent CLI. Three usage modes:
- Interactive TUI —
openmelon(no args, in a project) drops into a bubbletea REPL with slash commands, palette, model + skill pickers, bash approval modal, session resume. - Headless prompt —
openmelon -p "<intent>". Inside an OpenMelon project, runs the same tool-driven runtime as the TUI without the TUI. Outside a project, falls back to the legacy one-shot skillplus → LLM JSON → optional image/artifact path. Used by integrations and scripts. - Public Go surface —
pkg/openmeloncurrently exposes version metadata;pkg/contractsholds public contract types for embedding/integration surfaces.
Repo: https://github.com/eight-acres-lab/openmelon Module:
github.com/eight-acres-lab/openmelon
Layout
cmd/openmelon/
main.go subcommand dispatch + legacy flag-based one-shot
cmd_init.go openmelon init
cmd_project.go project list|use|show|set-key|unset-key|keys
cmd_registry.go character / reference / material add|list|show|rm
cmd_search.go openmelon search
cmd_repl.go runRepl: builds runtime + dispatches TUI vs bufio
cmd_resume.go openmelon resume [<id>]
cmd_setup.go openmelon setup (re-run auth wizard)
agent_runtime.go headless `-p` path inside a project
publish.go --publish vbox helper
internal/
userconfig/ ~/.openmelon/{config,credentials,projects}.json
+ IsTrusted / ResolveAPIKey (project → global → env)
projectx/ <workdir>/.openmelon/project.json + Settings + .gitignore
registry/ characters / references / materials on-disk store
(.search files, source-of-truth for description + tags)
search/ tag + grep, no vectors
llm/ pluggable Client (Complete/Stream) + ToolCaller (Chat)
+ StreamingToolCaller (StreamChat) + Usage tracking
imagegen/ pluggable Generator with ReferenceImages support
+ retry on transient TLS / 5xx + DisableKeepAlives
tools/ tool registry + builtin tools (list_characters,
get_character, search, compile_skill, generate_image,
save_artifact, bash, finish, ...)
+ bash judge LLM + per-session allowlist
runtime/ tool-using agent loop driven by llm.ToolCaller
+ Tracer interface + History support for multi-turn
session/ per-run messages.jsonl + meta.json + Recent / LoadHistory
onboard/ first-run wizards (trust → auth → project init) as
ONE alt-screen bubbletea program with state machine
+ providers.go (public Provider / Preset for /model)
tui/ bubbletea TUI: model.go (state machine), tui.go (entry),
tracer.go (runtime → tea.Msg bridge), style.go, keys.go,
messages.go (per-event tea.Msg types)
repl/ bufio fallback REPL for non-tty contexts (CI, pipes)
skillplus/ subprocess wrapper to the `skillplus` CLI + ListSkills
agent/ legacy 0.2 one-shot agent (used outside a project)
artifacts/ legacy artifact write helper
provenance/ legacy provenance JSONL helper
project/ legacy 0.1 project.json loader
workflow/ legacy 0.1 declarative workflow runner
generation/ legacy 0.1 generation providers (shell + LLM-backed adapter)
version/
pkg/
contracts/ public Go types
openmelon/ public version metadata (Version constant lives here)
npm/ @e8s/openmelon Node distribution (downloads the binary)
examples/
food-exploration/ legacy 0.1 declarative example
integrations/ Skill files for Claude Code, Cursor
assets/ logo + provenance
docs/ design notes, testing recipe
scripts/release.sh tag + build × 4 platforms + GH release + npm publish
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.
- 5d ago First seen · 136 lines · 2,113 tokens per session scan A cd6fac173da9
openmelon CLAUDE.md is an instructions file published in the GitHub repository eight-acres-lab/openmelon (87 stars, last pushed 3mo ago), licensed Apache-2.0. It adds 2,113 tokens to every session, about $0.0106 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.
Other instructions, from other repositories
tongflow CLAUDE.md
Claude Code instructions for tong-io/tongflow, covering claude.md, directory conventions, contract enforcement: compile-time only, plugin authoring rules and registering an official plugin.
Inline-Studio CLAUDE.md
Claude Code instructions for inlineresearch/Inline-Studio, covering inline studio - engineering guide, mental model (everything is organised around this), architecture, directory map (ts side) and storage - a portable project folder (owned by core).
plano CLAUDE.md
Instructions for katanemo/plano, covering claude.md, build & test commands, rust — wasm plugins (must target wasm32-wasip1), rust — brightstaff binary (native target) and rust — tests, format, lint.
stride-gpt AGENTS.md
AGENTS.md instructions for mrwadams/stride-gpt, covering agents.md, what this project is, repository layout, how the agentic analysis works and progressive disclosure pattern.
langchain-aws AGENTS.md
AGENTS.md instructions for langchain-ai/langchain-aws, covering global development guidelines for the langchain monorepo, project architecture and context, langchain-aws structure, key components and langgraph-checkpoint-aws structure.
Sponsio CLAUDE.md
Claude Code instructions for SponsioLabs/Sponsio, covering agent guide for sponsio, what sponsio is, positioning, what to read first and repository map.