agentleFS
s-morgan-jeffriess-morgan-jeffries/omnifocus-mcp/.claude/skills/applescript-omnifocus/SKILL.md

applescript-omnifocus

Skill8 starsChanged 8 months agoMITView on GitHub

Use when writing, modifying, or debugging ANY AppleScript that interacts with OmniFocus. Also use when encountering AppleScript errors, unexpected OmniFocus behavior, or when adding new connector methods. Covers variable naming traps, rich text limitations, recurring task handling, date formats, string escaping, and batch operation patterns.

The file

We split this skill into the rules it gives an agent

4 rules. About 0 are common, 3 are distinctive, and 1 is very specific to this project.

  • ≈0

    Common · estimate

    Many files like this have a rule of this kind.

  • ≈3

    Distinctive · estimate

    This author's own choices. Worth a look.

  • 1

    Very specific

    Tied to their project's names, files or tools. Probably skip.

  • Common
  • Distinctive
  • Very specific
  1. Distinctive:

    Safe naming pattern: Always prefix with the entity type:

    line 17
  2. Very specific:

    Use mark complete (the AppleScript command) for any task that might recur.

    line 51
  3. Distinctive:

    When writing new date-handling code, always use the existing converter rather than building date strings manually.

    line 87
  4. Distinctive:

    Always escape user-provided strings before embedding in AppleScript:

    line 91

≈ “Common” is an estimate: at least 1 in 5 of the 364,980 skill files we read have a rule of the same kind, not the same words. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.

Keep what fits and drop the rest. Your version, ready for your agents.
Files
SKILL.md

AppleScript + OmniFocus Patterns

This skill covers the hard-won knowledge about scripting OmniFocus via AppleScript. These are not obvious from documentation and have caused real bugs in this project.

Critical: Variable Naming Conflicts

OmniFocus AppleScript has a dangerous quirk: if you use a variable name that matches an OmniFocus property, AppleScript silently reads the property instead of your variable.

Bug-causing names (NEVER use as variables): name, note, status, flagged, completed, sequential, due date, defer date, estimated minutes, repetition rule, repetition method

Safe naming pattern: Always prefix with the entity type:

-- BAD: "name" collides with OmniFocus property
set name to "My Task"
set name of theTask to name  -- reads property of theTask, not your variable!

-- GOOD: prefixed variable name
set taskName to "My Task"
set name of theTask to taskName  -- works correctly

Real bug from this project: A variable called status inside a tell task block caused OmniFocus to read the task's status property instead of the variable value, producing silent data corruption. The elifintervalDays typo story: unit tests passed because they mocked AppleScript, but integration tests against real OmniFocus caught the variable collision. This is why integration tests exist.

Rich Text Notes Limitation

OmniFocus stores notes as rich text internally. The AppleScript API exposes this inconsistently:

  • Reading: note of task returns plain text (stripped of formatting)
  • Writing: set note of task to "text" replaces the entire note with plain text, destroying any rich text formatting

There is no workaround. You cannot preserve rich text through the AppleScript API. Document this to users and warn when overwriting notes.

Recurring Task Completion

Two ways to complete a task — only one handles recurrence correctly:

-- CORRECT: Creates next occurrence for recurring tasks
mark complete theTask

-- WRONG: Just marks this instance done, no next occurrence spawned
set completed of theTask to true

Use mark complete (the AppleScript command) for any task that might recur. Use set completed to true only when you specifically want to kill a recurring task's future occurrences.

The connector uses mark complete by default in update_task() when completed=True is passed.

JSON Construction in AppleScript

AppleScript has no native JSON support and no module system. Every AppleScript block that needs to return structured data must include inline JSON helper functions.

Current pattern (used 9x in the connector):

on escapeJSON(theText)
    set resultText to ""
    repeat with i from 1 to count of theText
        set c to character i of theText
        if c is "\"" then
            set resultText to resultText & "\\\""
        else if c is "\\" then
            set resultText to resultText & "\\\\\\\\"
        else
            set resultText to resultText & c
        end if
    end repeat
    return resultText
end escapeJSON

This duplication is intentional and necessary. Do not try to "DRY" it — AppleScript blocks must be self-contained because osascript -e executes each block independently.

Date Format Conversion

AppleScript uses locale-dependent date strings. The connector normalizes to ISO 8601:

  • AppleScript format: "March 5, 2026 5:00:00 PM"
  • ISO 8601 format: "2026-03-05T17:00:00"
  • Conversion: _iso_to_applescript_date() in the connector handles both directions

When writing new date-handling code, always use the existing converter rather than building date strings manually.

String Escaping

Always escape user-provided strings before embedding in AppleScript:

name_escaped = self._escape_applescript_string(name)
script = f'set name of theTask to "{name_escaped}"'

The _escape_applescript_string() method handles quotes and backslashes. Without this, a task name containing " will break the AppleScript block silently.

Batch Operation Patterns

Single AppleScript call for multiple items is always faster than multiple calls:

-- GOOD: One call, N items (100-300ms total)
tell application "OmniFocus"
    tell front document
        repeat with t in flattened tasks
            -- process all tasks in one block
        end repeat
    end tell
end tell

-- BAD: N calls, one item each (100-300ms PER call)
for task_id in task_ids:
    script = f'tell application "OmniFocus" ... get task {task_id} ...'
    run_applescript(script)  # 100-300ms each!

When batch is not possible: Some operations (like mark complete) must be done per-task because they trigger OmniFocus-internal workflows. In these cases, accept the overhead.

Error Handling Pattern

AppleScript errors should be caught in the script and returned as prefixed strings:

try
    -- operation
    return "SUCCESS"
on error errMsg
    return "ERROR: " & errMsg
end try

The Python side then checks for the prefix:

result = run_applescript(script)
if result.startswith("ERROR:"):
    raise Exception(result)

This is more reliable than catching subprocess.CalledProcessError, which doesn't always contain useful error messages from OmniFocus.

Database Safety System

The connector has a safety system that prevents destructive operations on production data:

  • Requires OMNIFOCUS_TEST_MODE=true and OMNIFOCUS_TEST_DATABASE=OmniFocus-TEST.ofocus
  • Each destructive operation calls _verify_database_safety() which runs an AppleScript to check the current database name
  • Operations tracked in DESTRUCTIVE_OPERATIONS set: create, update, delete, reorder

This adds ~100ms per destructive call but prevents catastrophic data loss. Do not bypass it.

Integration Test Requirement

If you wrote or modified AppleScript in the connector, you MUST write or update integration tests before the PR is ready. Unit tests mock run_applescript() — they prove your Python logic works but say nothing about whether the AppleScript is valid. Offer to run make test-integration before creating the PR.

OmniFocus-Specific AppleScript Quirks

  • first flattened task whose id is X — works for tasks. Use flattened projects for projects, folders for folders.
  • front document vs default document — front document is for UI operations, default document for data operations. Some operations require one or the other.
  • Focus — only projects and folders can be focused. Attempting to focus a task or tag raises an error.
  • Perspectives — perspective names are accessed via perspective names (plural), switching via set perspective name of window to X.
  • Move operations — move task to before/after referenceTask requires both tasks to be in the same project and at the same hierarchy level.

Download now (ZIP)

Does it work?

Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.

No reviews yet

People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.

  • It worked
  • Partly
  • It didn't
What did you use it for? e.g. Reviewing pull requests in a Rails app

Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.

Your trail

Files you open will collect here, so you can find them again.

Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.

Now do this for your own.

Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.