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/lcharvol/macsift/claude-mdgit clone --depth 1 https://github.com/Lcharvol/MacSiftWrote 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/lcharvol/macsift/claude-md)<a href="https://agentmods.dev/instructions/lcharvol/macsift/claude-md"><img src="https://agentmods.dev/badge/instructions/lcharvol/macsift/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 | $0.03640 | $0.03640 |
| Opus 5 | $0.01820 | $0.01820 |
| Sonnet 5 | $0.00728 | $0.00728 |
| Haiku 4.5 | $0.00364 | $0.00364 |
Grade B, and why
MacSift 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 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.
Asks for rootmediumPrivilege escalation
A mod that escalates privileges can change anything on the machine, not only the project.
- Time Machine snapshot deletion via `tmutil deletelocalsnapshots` may require admin privileges. The cleaning report adds a hint pointing the user to run `sudo tmutil deletelocalsnapshots <date>` in Terminal when this ha How it starts
The opening of the file, as written. The whole thing — 187 lines — stays where its author put it; the contents beside it link to each section on GitHub.
MacSift — Working Notes for Claude
Transparent macOS disk cleaning utility. SwiftUI, MVVM, built with SwiftPM (no Xcode project file). Deployment target is macOS 15 (Sequoia); Liquid Glass APIs from macOS 26 (Tahoe) are gated behind runtime availability checks via the compatGlassEffect / compatGlassButtonStyle helpers in Utilities/GlassCompat.swift.
Build / run / test
# Compile only (fast, for type-checking)
swift build
# Run the full test suite
swift test
# Run a single suite
swift test --filter CleaningEngineIntegrationTests
# Build the .app bundle and launch (this is how the user runs it)
./build-app.sh
open MacSift.app
build-app.sh produces a proper .app bundle (with Info.plist + AppIcon.icns) from the SPM build output, ad-hoc signs it for stable TCC identity, and prints next steps. The bundle path is MacSift.app/ at the repo root and is gitignored.
The user is on macOS 26.2 / Xcode 26.3. Deployment target is macOS 15.0 (Sequoia). On Sequoia the app falls back to .regularMaterial / .bordered instead of Liquid Glass.
Project layout
MacSift/
├── App/ # MacSiftApp entry point + AppState (@Published mode, dryRun, threshold)
├── Models/ # FileCategory, ScannedFile, ScanResult, FileGroup — pure value types, Sendable
├── Services/ # DiskScanner, CategoryClassifier, FileGrouper, CleaningEngine,
│ TimeMachineService, ExclusionManager
├── ViewModels/ # ScanViewModel, CleaningViewModel — @MainActor, @Published, orchestrate services
├── Views/ # SwiftUI views — WelcomeView, ScanProgressView, FileListSection,
│ FileGroupRow, InspectorView, MainView, …
└── Utilities/ # FileSize+Formatting, FileDescriptions, BundleNames, Permissions
Keep the structure flat. Don't introduce package boundaries or sub-frameworks.
File grouping architecture
ScannedFile is one underlying OS file. FileGroup is a display-layer
aggregation — multiple ScannedFiles that share a common owner (e.g., all files
under ~/Library/Caches/com.apple.Safari/ collapse into one "Safari" row).
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 · 187 lines · 3,640 tokens per session scan B 6c4cbbec7605
MacSift CLAUDE.md is an instructions file published in the GitHub repository Lcharvol/MacSift (269 stars, last pushed 1mo ago), licensed MIT. It adds 3,640 tokens to every session, about $0.0182 per session on Opus 5. A static security scan graded it B with 1 finding (asks for root). No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-30.
Other instructions, from other repositories
TokenBar AGENTS.md
AGENTS.md instructions for Nanako0129/TokenBar, covering tokenbar agent routing, read order, invariants, authorization boundary and handoff.
TokenBar CLAUDE.md
Claude Code instructions for Nanako0129/TokenBar: Read AGENTS.md first, then docs/knowledge/README.md and the task-specific canonical document it routes to. If .agent-local/CLAUDE.md exists, read it after the canonical documents as additive machine-local guidance only; it must not override the canonical architecture…
KumoApp AGENTS.md
AGENTS.md instructions for ProjectKumo/KumoApp, covering kumo agent guidelines, development documentation, documentation maintenance, ui copy constraints and swiftui native component constraints.
FluidAudio copilot-instructions.md
Copilot instructions for FluidInference/FluidAudio, covering fluidaudio - agent development guide, build & test commands, architecture, critical rules and code style (swift-format config).
speech-swift AGENTS.md
AGENTS.md instructions for soniqo/speech-swift, covering agent instructions, communication style, workflow, running tests and benchmarks — sequential and memory-aware and git conventions.
codex-provider-sync AGENTS.md
AGENTS.md instructions for Dailin521/codex-provider-sync, covering ai / agent operator guide, vnext architecture baseline, goal, choose the interface and safe operating flow.