macos-permissions
An agent skill by daymade, from daymade/claude-code-skills. Tags: access-control, debugging, macos, security, troubleshooting.
What it does
Diagnose and fix macOS TCC permission dialogs and silent denials — Screen Recording, Microphone, Camera, Accessibility, Automation (Apple Events), Full Disk Access, Files & Folders, and the "X would like to access data from other apps" prompt. Use whenever an app or background job is blocked by a macOS privacy permission, a permission dialog reappears after clicking Allow, a LaunchAgent/`uv run` job keeps prompting, System Settings shows a bare version number or wrong name, a granted permission silently stops working after an update, or you need to find WHO is really requesting a permission. Covers reading TCC.db as the ground-truth source, attribution (display name ≠ responsible process), per-binary-path grants, the `uv`-in-launchd Full-Disk-Access trap, `tccutil reset`, and SIP limits. 中文触发：权限弹窗、授权、TCC、完全磁盘访问、Full Disk Access、录屏/麦克风/摄像头/辅助功能/自动化权限被拒、访问其他应用的数据、弹窗一直弹、授权了没用、升级后失效、System Settings 里显示版本号。
Install
With the skills CLI, which installs into Claude Code, Codex, Cursor and other agents:
npx skills add daymade/claude-code-skills --skill macos-permissionsOr copy the skill folder into Claude Code's skills directory by hand (~/.claude/skills for every project, or .claude/skills inside one):
git clone --depth 1 https://github.com/daymade/claude-code-skills
cp -r claude-code-skills/daymade-macos/macos-permissions ~/.claude/skills/macos-permissionsSafety box score
Grade C, moderate risk: 42 of 100 risk points, rated 2026-10-20. The grade rates what the skill and its scripts can reach on the machine of whoever installs it, across eight categories from shell execution to secrets access. How the score works.
This skill diagnoses macOS TCC permission failures by reading TCC logs/databases, checking LaunchAgents, and steering the user through System Settings to add or reset grants. The main risks are privileged local access (FDA and sudo), changes to system privacy settings, and running shell diagnostics against protected paths; it does not fetch remote code or ship opaque executables.
| Category | Level | Why |
|---|---|---|
| Code & shell execution | 2 (moderate) | It directs the agent to run local commands such as `/usr/bin/log show`, `sqlite3`, `tccutil reset`, and sometimes `launchctl` for diagnosis and repair. |
| Network access | 0 (none) | There are no remote downloads, uploads, API calls, or network destinations in the skill text; the workflow is local macOS tooling only. |
| Filesystem reach | 2 (moderate) | It reads protected paths like `/Library/Application Support/com.apple.TCC/TCC.db`, `~/Library/Application Support/com.apple.TCC/TCC.db`, and `~/.local/bin/uv`, but does not describe destructive writes outside the project. |
| Secrets & credentials | 1 (low) | The skill may expose privacy-grant state and asks for system authentication in System Settings, but it does not request passwords, tokens, key files, or browser/CLI session secrets. |
| Privilege & persistence | 3 (high) | It explicitly requires `sudo` for the system TCC database and tells the user to grant Full Disk Access in System Settings, which is a broad machine-level privilege change. |
| Instruction hijack surface | 0 (none) | No hidden instructions, prompt-injection patterns, or requests to ignore higher-priority rules appear in the skill files. |
| Irreversible actions | 2 (moderate) | It can change Privacy & Security grants and reset TCC entries, but it repeatedly requires exact request verification and explicit authorization before any restart or relaunch. |
| Transparency | 0 (none) | The skill is plain markdown with readable commands and prose; there are no binaries, base64 blobs, or downloaded installers. |
Before installing
- Read `references/tcc-mechanics.md` first and confirm you are okay granting the terminal Full Disk Access and using `sudo` against `/Library/Application Support/com.apple.TCC/TCC.db`.
- Be ready for the skill to change macOS Privacy & Security settings and reset TCC entries; do not install it on a machine where that would be unacceptable.
- If you only need diagnosis, try it on a non-sensitive macOS account or test machine before using it on your main user profile.
Rated 2026-10-20 by an automated review of 4 files from the skill's repository.
Source
- Repository
- daymade/claude-code-skills (all skills from this repository)
- Path
- daymade-macos/macos-permissions/SKILL.md
- Branch
- main
- Updated
- 2026-09-20
Related skills
- tunnel-doctor — Diagnoses concrete tunnel and proxy-path failures across macOS and Windows/WSL: Tailscale routing, proxy env/system bypass, SSH double tunneling.
- address-sanitizer — Builds and runs code under AddressSanitizer to catch buffer overflows, use-after-free, and other memory errors during fuzzing or tests.
- agents-debug — Use when your agent or environment is broken — wrong answers, errors, timeouts, tool failures, or CLI issues. Reads traces and logs to diagnose root causes.
- agents-harden — Use when preparing your agent for production — IAM scoping, inbound auth (JWT, SigV4), secrets management, cold start optimization, session lifecycle.
- app-platform-troubleshooting — Debug running App Platform applications by accessing containers, analyzing logs, running diagnostics, and applying fixes.
- audit-context-building — Understand a codebase before looking for bugs in it - what each function assumes, what it guarantees, and what it depends on elsewhere.