project
launchd is Apple's system and service management daemon for macOS, launching and scheduling daemons and LaunchAgents via plist files and launchctl.
No evidence-backed relationships are recorded.
One developer's weekly Claude Code skill update failed on 6, 13 and 20 September despite three layered fixes and an audit note reading "exit 0". The script caught and logged every failure correctly, but the log had no reader until the author opened it after the third one.
Reality
Every 10 minutes an unattended job summarizes Claude Code and Codex logs into an Obsidian vault and pushes the commit, so the pipeline treats the summarizer's own JSON as text that may carry pasted keys or injected instructions.
Reality
A dev.to harness guide puts numbers on two ways MCP breaks at scale, 8,000 to 15,000 tokens of tool schema on every prompt and a SQLite session lock duplicate stdio processes cannot share. Its fix moves stateful tools onto a supervised loopback daemon.
Reality
Offset start times kept a shared file queue from colliding for months. One day-long Codex outage made every launchd lane start at the same instant, and the pickup scripts had no mutual exclusion to fall back on.
Reality
Stopping all 15 guarded jobs uniformly cut a posting job to a 43% execution rate, so the guard now lets a job flagged --priority send one real request every 1800 seconds, from a budget every priority job shares.
Reality
A strings trace of claude 2.1.266 puts the allowed-apps list in session memory with no key to pre-grant it, so one developer answered the dialog with a daemon that reads Apple Terminal tabs every two seconds.
Reality
Four Playwright jobs on one Mac held com.google.Chrome, so a Dock launch came back with error -600 while every job log stayed clean. The repair also had to survive npm install, which keeps resetting the browser path inside node_modules.
Reality
A prefix match against Instagram's rendered profile page missed the emoji the platform serves as an image, so a job that had already posted logged ok:false and the scheduler sent the carousel twice more, though the API said the post was there.
Reality
Over four days one daily publishing job failed three separate ways in the place it actually runs, and every one of those failures had already passed a hand-run test in the author's own interactive shell.
Reality
A macOS daemon held 47GB of mostly compressed memory while reporting 264MB of RSS, and the script that spotted it had no permission to act. The repair is one sudoers line for one killall, verified after install.
Reality
The per-session spawn is what stdio means, so the fix is one loopback daemon per machine. That same consolidation makes a single restart blind every open session, which is where the design work went.
Reality
Prompting an LLM to self-audit works some fraction of the time, so one developer moved the check into a Stop hook that parses the session transcript and refuses the end of a turn when the audit keywords are missing.
Reality
Execing the binary inside Google Chrome.app registers com.google.Chrome with LaunchServices. That registration outlives the script. The durable repair is a second bundle id, not a kill loop.
Reality
An adversarial self-audit check that blocks with exit 2 is a useful nag in a terminal. Under launchd it becomes a failed job. The fix here keys off session provenance rather than softening the rule.
Reality
The per-tab figures were summed off an unfiltered ps listing, so a browser policy that looked 198 MB ahead was really 91 MB behind. What caught it was watching one number fail to move across a page load.
Reality
Apple has been prying cron loose from the user session since Ventura. The result is a Sequoia Mac that can print every crontab entry with nothing behind it, so the honest test is whether launchctl will admit to a cron job at all.
Reality
Command substitution inside the echo reset the exit status to zero, so a Stop hook reported success for a full day while DM replies sat at zero on two platforms. The test that catches this takes a minute to write.
Reality
The kill pattern matched a Chrome profile path rather than a process the caller owned, so a posting job in one repository SIGKILLed a Playwright job in another and took the cookie database with it.
Reality
Fable 5 shuts down on 2026-07-07 and StartCalendarInterval cannot express a single firing at a fixed datetime, so this job leans on an integer date test, a temp-file jq rewrite, and an unload that runs whether or not the rewrite worked.
Reality
A daily automation job was killed for two weeks because two reasonable config values were never multiplied together. The check that catches it runs before the first action.
Reality
Build · August 26, 2026 · 1 publisher
Build · August 26, 2026 · 1 publisher
Build · August 24, 2026 · 1 publisher
Build · August 23, 2026 · 1 publisher
Build · August 23, 2026 · 1 publisher
Build · August 22, 2026 · 1 publisher
Build · August 22, 2026 · 1 publisher
Build · August 21, 2026 · 1 publisher
Build · August 20, 2026 · 1 publisher
Build · August 20, 2026 · 1 publisher
Build · August 20, 2026 · 1 publisher
Build · August 18, 2026 · 1 publisher
Build · August 18, 2026 · 1 publisher
Build · August 18, 2026 · 1 publisher
Build · August 17, 2026 · 1 publisher
Build · August 17, 2026 · 1 publisher
Build · August 16, 2026 · 1 publisher
Build · August 15, 2026 · 1 publisher
Build · August 15, 2026 · 1 publisher