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/kaandedeoglu/shark/claude-mdgit clone --depth 1 https://github.com/kaandedeoglu/SharkWhat 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 | $0.02803 | $0.02803 |
| Opus 5 | $0.01401 | $0.01401 |
| Sonnet 5 | $0.00561 | $0.00561 |
| Haiku 4.5 | $0.00280 | $0.00280 |
Grade A, and why
Shark 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 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 — 188 lines — stays where its author put it; the contents beside it link to each section on GitHub.
CLAUDE.md
This file provides guidance to Claude Code (claude.ai/code) when working with code in this repository.
Overview
Shark is a Swift command line tool with three subcommands:
generate(default) — type-safe enums for iOS/macOS application assets (images, colors, fonts, localizations, storyboards, data assets). Parses the Xcode project to discover resources; compile-time safety for asset access.lint— localization completeness checks across all locales (missing-key, empty-source-value, placeholder-mismatch, orphaned-key); exit code 1 on findings, built as a CI gate.translate— fills localization gaps via local Claude Code by default, with Claude API and Codex CLI backends also supported; machine-validated format-specifier preservation andneeds_reviewwrite-back.
Positioning (keep docs/marketing consistent with this)
Shark's pitch against Xcode's generated symbols (which sherlocked images/colors in Xcode 15 and string accessors via String Catalogs later): Shark also covers fonts, storyboards, and data assets; speaks classic .strings and .xcstrings; namespaces by folder structure and key separators; handles multi-target/white-label setups (--target, --exclude, --name); and — the core argument — treats localization as a workflow (generate → lint → translate), not just codegen. Against SwiftGen/R.swift: project-driven instead of config-driven, zero runtime dependency, and they don't lint or translate. XcodeGen/Tuist are project generators — orthogonal, not competitors. The three subcommands deliberately share one project model and one FormatSpecifierParser, so claims like "lint checks exactly what generate generates" stay true — preserve that property when extending.
Build and Development Commands
Building the Project
# Build in debug mode
swift build
# Build for release
swift build -c release
# Install locally (after building release)
cp ./.build/release/Shark /usr/local/bin/shark
Running Tests
The project uses Swift Package Manager's built-in testing. Look for test files in the standard Tests/ directory structure.
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 · 188 lines · 2,803 tokens per session scan A 32fc25a547ab
Shark CLAUDE.md is an instructions file published in the GitHub repository kaandedeoglu/Shark (382 stars, last pushed 11d ago), licensed MIT. It adds 2,803 tokens to every session, about $0.0140 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
translate-book CLAUDE.md
Instructions for deusyu/translate-book, covering claude.md, project, structure, testing changes and then run translation via the skill.
hana-developer-cli-tool-example translatable-text-handling.instructions.md
Use when adding user-facing text in CLI commands, API routes, or utilities. Enforces consistent use of bundle.getText() for all translatable strings, descriptive key naming, and multi-language synchronization across i18n/ property files.
ai-dial-chat rtl.instructions.md
Instructions for epam/ai-dial-chat, a project described as: A default UI for AI DIAL.
bob-plugin-ollama-translator AGENTS.md
Instructions for CaicoLeung/bob-plugin-ollama-translator, covering repository guidelines, project overview, architecture & data flow, key directories and development commands.
xcstrings-crud CLAUDE.md
Instructions for Ryu0118/xcstrings-crud, covering xcstrings-crud, build & test, architecture, code style and key files.
next-app-i18n-starter AGENTS.md
Instructions for s0vers/next-app-i18n-starter, covering agent guide, project, non-negotiables, cursor rules and deep reference.