brianharms/reminder-watch/skill
Claude Code and agent skill by brianharms/reminder-watch
- community
Checking that the SKILL.md file is still in the repository.
Check and act on Apple Reminders — reads Claude Inbox for pending tasks and Claude Output for results. Automatically dispatches agents for new inbox items. Use when user says /reminders, "check reminders", "what came in", or to start the watcher via /loop.
Install
npx vskill@latest install brianharms/reminder-watch/skillSKILL.md
Reminders Bridge
Check Apple Reminders for new tasks and dispatch agents to work on them.
Step 1: Config Check
Check if ~/.claude/reminders-config.sh exists.
If it exists: Source it to get the PROJECTS_DIR variable. Continue to Step 2.
If it does NOT exist:
- First, check if this skill was invoked via
/loop. If so, print this message and stop:
> Run /reminders once to configure your projects folder, then start the loop.
- If invoked directly, ask the user: "Where's your projects folder?"
- Accept natural language (e.g., "it's in myProjects on the desktop", "~/code", "the projects folder on my desktop")
- Resolve the response to an absolute path (expand
~, interpret natural language like "on the desktop" →~/Desktop/...) - Verify the path exists by running
lson it - If the path doesn't exist, tell the user and ask again
- Once valid, write the config file:
``bash echo 'PROJECTS_DIR="/resolved/absolute/path"' > ~/.claude/reminders-config.sh ``
- Continue to Step 2
Step 2: Run the Check Script
Run this exact command:
This script:
- Queries "Claude Inbox" for incomplete reminders
- Filters out already-seen IDs (tracked in
~/.claude/reminder-seen.txt) - Marks new reminders as complete in Reminders (prevents re-processing)
- Queries "Claude Output" for flagged items (questions needing user input)
- Outputs pipe-delimited lines:
NEW|title|notesorFLAGGED|title|notes - Returns exit code 1 if nothing new (no output)
If exit code is 1 (nothing new): Say nothing. Return silently. Do NOT output "nothing new" or any status message — just stop.
Step 3: Handle Results
For each `NEW|title|notes` line:
- Run
lson thePROJECTS_DIRto get the list of existing project folder names. - Evaluate the task title and notes against the folder names. Use your judgment:
- Clear match — Use that folder (e.g., "work on mac caps" →
mac-caps/) - Ambiguous but close — Lean toward the existing project when the name is similar
- No match — Create a new folder in
PROJECTS_DIRwith a concise kebab-case name derived from the task title
- Dispatch a background Agent:
For each `FLAGGED|title|notes` line: Display it prominently — these are questions from agents that need user input.
Step 4: Report
- If agents were dispatched: "Dispatched agent for: [title] → [project-folder]" (one line per task)
- If flagged items exist: "Needs your input: [title]"
- If nothing at all: say nothing (silent return)
Security
- Security scanPassed, 100 of 100Pattern checks on the skill's files.
- Model reviewPassed
Quality evaluation
- Task evaluationNot runNo evaluation yet. Run one in Skill Studio.
Source
- SKILL.md in the repositoryCheckingskill/SKILL.md
Popularity
- Repository stars
- 57
- Repository forks
- 5
- 7-day trend score
- 20
Stars and forks are for the whole GitHub repository, not this skill alone. Refreshed 29 Apr 2026, 6 months ago.
Works with
Any agent vskill installs to: 56 targets in all. These 39 have a page on this site.
- Claude Code
- Cursor
- GitHub Copilot
- Windsurf
- Codex
- Gemini CLI
- Cline
- Amp
- Roo Code
- Goose
- Continue
- Kilo Code
- Junie
- Trae
- OpenHands
- Qwen Code
- AI Agent
- Tabnine Chat
- Sourcegraph Cody
- Amazon Q
- Firebase Genkit
- Sweep
- Devin
- Phind
- Mentat
- Claude Desktop
- Aider
- GPT Pilot
- Plandex
- bolt.new
- v0
- Replit Agent
- Anse
- TypingMind
- BetterChatGPT
- LibreChat
- Warp AI
- Pieces
- Fig AI
Embed badge
[](https://verified-skill.com/skills/brianharms/reminder-watch/skill)