Install
$ agentstack add skill-ozlar34-claude-code-skills-claude-inbox ✓ scanned · ✓ verified, works with Claude Code, Cursor, and more.
Security review
✓ PassedNo issues found. Passed automated security review. · v0.1.0 How review works →
- ✓ Prompt-injection patterns
- ✓ Secret / credential exfiltration
- ✓ Dangerous shell & filesystem operations
- ✓ Untrusted network calls
- ✓ Known-malicious package signatures
What it can access
- ✓ Network access No
- ✓ Filesystem access No
- ✓ Shell / process execution No
- ✓ Environment & secrets No
- ✓ Dynamic code execution No
From automated source analysis of v0.1.0. “Used” means the capability is present in the source — more access means more to trust, not that it’s unsafe.
Verified badge
Passed review? Show it. Paste this badge into your README, it links to the public security report.
Reliability & compatibility
Declared compatibility
Compatibility is declared by the source manifest. End-to-end runtime verification is coming, see below.
We're building live execution health for every listing: tool-call success rate, median latency, uptime, and last-checked timestamps, measured, not self-reported. It isn't live yet, so we don't show numbers we can't stand behind.
How agent discovery & health will work →About
Claude Inbox — Mobile Capture Processor
Process ideas captured on mobile via Siri → Apple Reminders → an iCloud-synced list. GTD-style: one at a time, analyze, decide, execute, complete.
Configuration
Edit these defaults at the top of this file before first use:
LIST_NAME : Claude Inbox
ROUTE_TO_TASK_TOOL :
ROUTE_TO_NOTES :
The list name must match exactly what you create in Apple Reminders. The two routing destinations are descriptive — Claude will use whatever tools/MCPs you have wired up.
How items get here
You say: "Hey Siri, add [idea] to " from your phone or watch while away from the Mac. Items sync via iCloud Reminders.
Reading the list (AppleScript)
osascript "
set activeItems to (reminders of inboxList whose completed is false)
repeat with r in activeItems
set output to output & (id of r) & "|||" & (name of r) & "|||" & ((creation date of r) as string)
if body of r is not missing value then
set output to output & "|||" & (body of r)
end if
set output to output & linefeed
end repeat
return output
end tell
APPLESCRIPT
Replace `` literally before running. If the list doesn't exist, AppleScript errors. Tell the user to create a Reminders list with the exact name they configured and try again.
Marking an item complete
osascript -e 'tell application "Reminders" to set completed of (first reminder whose id is "REMINDER_ID") to true'
Use the id captured from the read step.
Workflow
Step 0: Load inbox
- Read the list. Count items. Report: "You have N items in ."
- If empty, say so and stop.
Step 1: Present one item
For each reminder:
### [N/total] Title
**Captured:** date (relative — "2 hours ago", "yesterday")
**Notes:** [body if present]
**Interpretation:** What you think the user meant by this. One sentence.
**Relevance:** How it connects to active projects/goals — or honestly, that it doesn't.
**Recommended action:** [one below] — because [reason]
Step 2: User decides
| Action | What happens | |--------|-------------| | Act now | Do it in this session if small + bounded; for anything bigger, suggest creating a planning artifact instead. | | Task | Create a task in the configured ROUTE_TO_TASK_TOOL (confirm title/project/priority/date before creating). | | Note | Append to or create a note in ROUTE_TO_NOTES. Ask where if ambiguous. | | Backlog | One-liner to a "tool/idea backlog" file in ROUTE_TO_NOTES. | | Ignore | Drop it — bad idea in hindsight, already done, duplicate. |
Step 3: Execute and complete
- Execute the chosen action.
- Ask: "Mark this item complete in ?" — wait for confirmation.
- On confirmation, mark complete via AppleScript using the item's id.
- Move to the next item.
Step 4: Wrap up
### processed: N
- Acted: X
- Tasks created: X
- Notes saved: X
- Backlog: X
- Ignored: X
Rules
- One at a time. Never batch.
- Be honest about relevance. If an item doesn't connect to anything, say so.
- Voice-captured text is noisy. Interpret generously — Siri mishears words. If ambiguous, ask what was meant before routing.
- "Stop" ends the session. Still show the summary.
- Complete means complete. Don't delete reminders — marking complete preserves history and keeps the list clean.
Source & license
This open-source skill is cataloged on AgentStack and links to its original source — we do not rehost the code.
- Author: ozlar34
- Source: ozlar34/claude-code-skills
- License: MIT
- Homepage: https://github.com/ozlar34
Install and usage instructions live in the source repository linked above.
Reviews
No reviews yet, be the first.
Write a review
Versions
- v0.1.0 Imported from the upstream source.