Apple Calendar Ops
This skill is the Apple Calendar operation layer.
It handles stable calendar reads and writes. It does not do high-level scheduling, cross-system planning, or task prioritization.
Core boundary
Use this skill for concrete Apple Calendar operations:
- list calendars
- fetch events in a time range
- create an event
- update an event
- delete an event
Do not use this skill to decide how the day should be planned. That belongs to a higher-level task/orchestrator.
Operating rules
Default stance:
- reads are safe
- writes should be dry-run-friendly
- updates/deletes should prefer explicit event ids
- fuzzy title matching may help locate events, but should not be the only basis for risky writes
Credentials should come from /home/agent/.openclaw/workspace/secrets.json.
Read references/boundary.md before changing the skill's scope.
Read references/event-contract.md before writing or consuming event JSON.
Quick start
Read calendars
python3 /home/agent/.openclaw/workspace/skills/apple-calendar-ops/scripts/calendar_fetch.py --list-calendars
Read events
python3 /home/agent/.openclaw/workspace/skills/apple-calendar-ops/scripts/calendar_fetch.py \
--start 2026-03-12T00:00:00+08:00 \
--end 2026-03-13T00:00:00+08:00
Create event
python3 /home/agent/.openclaw/workspace/skills/apple-calendar-ops/scripts/calendar_create.py \
--calendar "Calendar" \
--title "Example event" \
--start 2026-03-12T14:00:00+08:00 \
--end 2026-03-12T15:00:00+08:00 \
--dry-run
Scripts
scripts/calendar_common.py— shared config, secret loading, and JSON helpersscripts/calendar_fetch.py— list calendars and fetch eventsscripts/calendar_create.py— create an eventscripts/calendar_update.py— update an eventscripts/calendar_delete.py— delete an event
References
references/boundary.md— scope and non-goalsreferences/event-contract.md— normalized event shape for all scripts
First-version goal
Version 1 should make Apple Calendar readable and safely writable.
That means:
- reliable read access for scheduler inputs
- explicit create/update/delete flows
- machine-readable output
- conservative handling of risky writes