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/palamim/starboard/claude-mdgit clone --depth 1 https://github.com/palamim/starboardWrote 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/palamim/starboard/claude-md)<a href="https://agentmods.dev/instructions/palamim/starboard/claude-md"><img src="https://agentmods.dev/badge/instructions/palamim/starboard/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 5.1 | $0.01922 | $0.01922 |
| Opus 5 | $0.00961 | $0.00961 |
| Sonnet 5 | $0.00384 | $0.00384 |
| Haiku 4.5 | $0.00192 | $0.00192 |
Grade A, and why
starboard 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 today.
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 — 139 lines — stays where its author put it; the contents beside it link to each section on GitHub.
CLAUDE.md
Guidance for Claude Code working in this repo.
Commands
- Build:
swift build - Run:
.build/debug/Starboard, backgrounded (swift runblocks the terminal on stdout/stderr) - No tests, linters, or CI beyond
.github/workflows/build.yml.
Code style
Zero comments in Sources/Starboard/, by design — this file is the only
place rationale lives. If you add non-obvious logic, document the why
here, not inline.
Architecture
Plain SPM executable, no Xcode project, no Info.plist — main.swift sets
.accessory activation policy directly instead. AppDelegate's behavior
is split across AppDelegate+*.swift extensions by concern (the invisible
main menu bar, the clickable popup panel menu, Dock-tracking state
machine, Dock-relative geometry, screen resolution, Dock Accessibility
reads, debug logging, theme switching, panel appearance settings, the
Accessibility fallback hint); DockPresence, PanelBuilder,
TerminalTheme, PanelSettings, Theme, ThemePickerView,
SettingsPanelView, TerminalLayout, ShellEnvironment,
FallbackHintPanel, and Mascot are standalone. Read the code for how it
works — it's small and the names are literal.
Panel corner radius, background tint opacity, and terminal font are all
user-adjustable at runtime via the "Settings…" panel menu item
(SettingsPanelView/AppDelegate+Settings.swift), persisted in
UserDefaults through PanelSettings. Corner radius always has a
concrete value (defaults to TerminalTheme.defaultCornerRadius until
overridden); tint opacity is nil until the user touches the slider, so
switching themes still shows each theme's own baked-in alpha rather than
silently forcing one opacity on every theme; font name likewise falls
back to TerminalTheme.defaultFontName (the first installed preferred
Nerd Font, else the system monospace font) until overridden. The font
picker enumerates installed fonts, keeps those marked monospaced by
AppKit, and presents one regular-preferred face per family while storing
its PostScript name. Panel width was deliberately left out of this — the whole
product idea is gluing to the Dock's tray width, so that one stays a
hardcoded constant in TerminalLayout.
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.
- today Changed · +2 lines · +30 tokens per session e3ccd3c232b1
- 5d ago First seen · 137 lines · 1,892 tokens per session scan A 9aeb8c745e96
starboard CLAUDE.md is an instructions file published in the GitHub repository palamim/starboard (339 stars, last pushed yesterday), licensed MIT. It adds 1,922 tokens to every session, about $0.0096 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
swift-tui AGENTS.md
AGENTS.md instructions for SwiftTUI/swift-tui, covering agents.md, build & test commands, architecture (one-page summary), development guidelines and code style.
apple-dev-skills AGENTS.md
Instructions for gaelic-ghost/apple-dev-skills, covering agents.md, scope, apple rules, install guidance and validation.
FluidAudio CLAUDE.md
Claude Code instructions for FluidInference/FluidAudio, covering claude.md, project overview, critical development rules, never use @unchecked sendable and never create dummy models or synthetic data.
TMDb CLAUDE.md
Claude Code instructions for adamayoung/TMDb, covering claude.md, project overview, topic docs, knowledge base and development workflow.
liney AGENTS.md
AGENTS.md instructions for everettjf/liney, covering liney repository collaboration guide, 交流语言 / communication language, project overview, repository layout and build and test.
MarkdownView AGENTS.md
Instructions for keitaoouchi/MarkdownView, covering agents, technical components, relationships (data/event flow overview), extension points (overview) and platform / distribution.