Borrowing it
Nothing to install: this file belongs to sorkila/lockpaw. Take a copy, put it at the same path in your own repository, and replace the rules that are about this project with yours.
curl -O https://raw.githubusercontent.com/sorkila/lockpaw/main/CLAUDE.mdgit clone --depth 1 https://github.com/sorkila/lockpawWrote 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/sorkila/lockpaw/claude-md)<a href="https://agentmods.dev/instructions/sorkila/lockpaw/claude-md"><img src="https://agentmods.dev/badge/instructions/sorkila/lockpaw/claude-md/github.svg" alt="Measured on agentmods" height="20"></a>Or the 80×15 button, for a site that already has a row of RSS and ATOM ones. Only the verdict fits; the numbers stay here.
<a href="https://agentmods.dev/instructions/sorkila/lockpaw/claude-md"><img src="https://agentmods.dev/badge/instructions/sorkila/lockpaw/claude-md.svg" alt="Reviewed on agentmods" width="80" 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 | $0.21426 | $0.21426 |
| Opus 5.5 | $0.08570 | $0.08570 |
| Sonnet 5.5 | $0.04285 | $0.04285 |
| Haiku 4.5 | $0.02143 | $0.02143 |
Grade D, and why
lockpaw CLAUDE.md scanned grade D with 2 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 12d 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.
- **Six hookable agents (since v1.2.0)** — Claude (`Notification`+`Stop`+`StopFailure`, `~/.claude/settings.json`), Codex (`notify` must be a **top-level** TOML key: `AgentHookConfig.mergingCodexNotify` inserts it before Recursive force deletehighDestructive command
rm -rf with a variable or a broad path is one typo away from removing the wrong tree.
- **Latest release** — v1.7.1 released 2026-10-26 (build 26). DMG SHA-256: `59229c8eff9dfb7ba8cb73d06ad623a56e41f938130d620e94b5af25cc9c28fa` (13 931 976 bytes). The supporter ask only: card after 7 days and 5 unlocks (w How it starts
The opening of the file, as written. The whole thing — 389 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Lockpaw
macOS menu bar screen guard. Lock/unlock with a hotkey; the covered screen glows when your AI agent (Claude Code / Codex / Gemini) needs you. Dog or cat mascot.
Quick reference
- App name: Lockpaw
- Bundle ID:
com.eriknielsen.lockpaw - URL scheme:
lockpaw:// - Website: getlockpaw.com (hosted on Inleed, deployed via FTP from
sorkila/lockpaw-web) - Repo: [email protected]:sorkila/lockpaw.git
- Requires: macOS 14+, Xcode 16+, XcodeGen
- Dependencies: Sparkle (SPM, auto-updates with EdDSA signing)
- Current version: 1.7.0
- Size: ~14 MB DMG download, ~17 MB installed (2.8 MB of that is Sparkle; 2.4 MB is the 1.6 mascot art) — keep README/site/marketing claims in sync with the actual DMG when this changes
Build
xcodegen generate
xcodebuild -project Lockpaw.xcodeproj -scheme Lockpaw -configuration Debug build
After each rebuild, reset TCC (binary signature changes invalidate Accessibility permission):
tccutil reset Accessibility com.eriknielsen.lockpaw
Test
xcodebuild -project Lockpaw.xcodeproj -scheme Lockpaw -configuration Debug test
338 unit tests covering LockState transitions, Constants formatting, HotkeyConfig conflict detection/auth-required unlock preference/hotkey matching, SleepPreventer state handling, Mascot resolution (incl. the hidden and custom cases), CustomMascot store (atomic install, downsampling, rejection leaves the old image intact, cached hasImage), PingDecision agent-ping branching, AgentPing payload decoding (hook event → kind, agent/project names, CLI forwarding), AgentHookConfig hook merging (in-place upgrade, matcher, foreign hooks untouched), AgentConnection (which agents already have the hook, read from their configs), FadeToBlack preference resolution (checkbox × delay), every branch of the PresentationLogic fade-to-black reducer, TerminationPolicy (quit refused unless .unlocked), OverlayPolicy (no overlay is ever transparent to the pointer; only the primary takes key, except the clicked screen for one Mirror-mode auth attempt), LockdownPolicy (Space-switching gesture types incl. the Dock's control stream are swallowed, also while the auth dialog has the keyboard), ScreenLayout (overlays rebuild only on real display changes), PingGate (each session announced once per kind per lock), PingIntake (every ping recorded, only the announcement gated), AgentPingQueue (one row per session, the glow speaks for the most urgent outstanding ping, urgent-first rows, a row names who and not what), LidSleepPolicy/SystemLockPolicy (lid-closed cutouts; a forged macOS-unlock notification can't unlock), the helper's code-signing requirement, WebhookRelay (exactly what leaves the Mac) and its throttle, SupporterLicence/SupportAsk (the once-ever card's gates, the yearly menu line)/SeasonalSkin, the Codex TOML placement, that every bundled mascot asset ships, PassiveAuthPolicy (when the Touch ID sensor may be armed × how each LAError ending is scored, including that no ending can spend an unlock attempt), InputBlockPolicy (a held hotkey toggles once; the deferred keyboard-blocker start only runs while still locked), PasswordEntryPolicy (the kiosk options that hold the keyboard while the lock-screen password field is open, its idle close, and the Command-key filter), DesktopWallpaper (per-screen lookup in the private wallpaper store — this Space and display before global, leftovers skipped — and that malformed or older stores fail safe to the still), WallpaperPlaybackPolicy (when the shared Aerial player runs), and ImageDownsampling (wallpaper stills decoded at their fill size).
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.
- 12d ago Changed · +359 tokens per session 1c2e0a6a001c
- 17d ago Changed · +13 lines · +1,464 tokens per session a3526020e624
- 18d ago Changed · +28 lines · +3,040 tokens per session 4accc7cb108d
- 23d ago Changed · +3 lines · +783 tokens per session 1ad70d72a7c3
- 25d ago Changed · +40 tokens per session 58d49c3cfb68
- 1mo ago Changed · +49 lines · +4,059 tokens per session ee29ee212823
- 1mo ago Changed · +10 lines · +664 tokens per session 6e0bedcb29aa
- 1mo ago Changed · +12 lines · +1,765 tokens per session 163ffc99086f
- 2mo ago Changed · +2 lines · +220 tokens per session b797a5ee168f
- 2mo ago First seen · 272 lines · 9,032 tokens per session scan D f5e41f7cb02c
lockpaw CLAUDE.md is an instructions file published in the GitHub repository sorkila/lockpaw (174 stars, last pushed 12d ago), licensed MIT. It adds 21,426 tokens to every session, about $0.0857 per session on Opus 5.5. A static security scan graded it D with 2 findings (reads agent configuration directories, recursive force delete). No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-30.
lockpaw CLAUDE.md compared with instructions from other repositories
| Mod | Type | Stars | Last push | Licence | Scan | Tokens | Works with |
|---|---|---|---|---|---|---|---|
| lockpaw CLAUDE.mdsorkila/lockpaw | Instructions file | 174 | 12d ago | MIT | D | 21,426 | Claude Code, Codex |
| vscode buildNext.instructions.mdmicrosoft/vscode | Instructions file | 195k | today | MIT | A | 9,719 | GitHub Copilot |
| vscode oss-third-party-notices.instructions.mdmicrosoft/vscode | Instructions file | 195k | today | MIT | A | 5,516 | GitHub Copilot |
| next.js AGENTS.mdvercel/next.js | Instructions file | 143k | today | MIT | A | 7,461 | Codex, OpenCode |
| langchain AGENTS.mdlangchain-ai/langchain | Instructions file | 149k | today | MIT | A | 4,680 | Codex, OpenCode |
| vscode copilot-instructions.mdmicrosoft/vscode | Instructions file | 195k | today | MIT | A | 3,093 | GitHub Copilot |
| vscode model-prompts.instructions.mdmicrosoft/vscode | Instructions file | 195k | today | MIT | A | 2,022 | GitHub Copilot |