macOS Launch Agents are property list files that tell the operating system to start per-user background processes at login or on a schedule. They run in the user context, unlike Launch Daemons, and are managed by launchd through defined labels, paths, and triggers.
What macOS Launch Agents Do
macOS Launch Agents are a user-level autostart mechanism, so they matter because they define what runs when a person logs in and what runs in that person’s context. That makes them operationally important for convenience and automation, but also security-relevant whenever background code should be treated as part of a user session rather than a system service.
They are described by property list files and interpreted by launchd, which means the security question is not just whether a file exists, but whether its label, path, trigger and ownership are consistent with expected user activity. This is one reason persistence tradecraft on macOS often includes Launch Agents alongside other autostart mechanisms, as discussed in NHIMG’s Persistence Techniques Guide.
How Launch Agents Differ From Launch Daemons
The most important distinction is execution context. Launch Agents run for a logged-in user and typically participate in that user’s desktop session, while Launch Daemons are intended for system-wide background work outside the user session. That difference changes what the process can see, which files and settings it can touch, and how defenders should reason about persistence and visibility.
Because they are user-context processes, Launch Agents can interact with user data, preferences and application state in ways that system services cannot. That also means a malicious or unwanted Launch Agent can blend into normal login activity, especially if the file name and plist contents look routine or if the agent is installed as part of a legitimate-looking application.
For defenders, this distinction is not academic. NHIMG’s T1543 Create or Modify System Process Defensive Guide covers the broader persistence family that includes Launch Agents and helps place them in the same operational pattern as other autostart mechanisms. The macOS-specific angle is that user-level persistence can survive routine reboots and still remain tied to an active account session.
Common Legitimate Uses and Abuse Paths
Legitimate Launch Agents are common on macOS because many applications need to sync data, check for updates, maintain helper processes or trigger user-facing features after login. In that sense, they are a normal part of the platform rather than a sign of compromise.
The security issue is that the same mechanism can be used for persistence, privilege-adjacent user execution or hidden background activity. A Launch Agent can be planted through application bundles, post-install actions, user-writable locations or other execution paths that make the process look like expected software behaviour. Once registered, it can relaunch without further user interaction.
That is why macOS persistence investigations often pair Launch Agent review with broader telemetry such as login items, launchd metadata, file creation events and process ancestry. NHIMG’s macOS Detection and Response Guide is useful here because it places Launch Agents alongside the telemetry and hunting patterns defenders use to confirm whether a background process is expected or suspicious.
What Defenders Look For
Defenders usually focus on where the plist lives, who owns it, what executable it points to, and whether it is launching from an unusual path or with unusual timing. A user-context startup item is more suspicious when it appears unexpectedly, points to a mutable location, or is associated with software the organisation did not approve.
Reviewing Launch Agents also helps with post-compromise analysis. If a host has been accessed by malware or an infostealer, the presence of a new or modified agent can show how the attacker tried to keep code running across logins. In macOS incident response, that often turns into a hunt for adjacent persistence, credential theft and any process that was designed to survive user restarts.
NHIMG’s ATT&CK macOS Malware Software Hub is a useful reference point because it connects Launch Agents to the wider macOS malware ecosystem, including persistence and credential theft patterns defenders commonly have to review together.
Risk and Threat Considerations
Launch Agents are a durable persistence point because they execute automatically in a user session and can survive routine logouts and reboots. That makes them attractive for malware that wants repeat access, stealthy background execution or access to user-level data without needing a system service.
Failure mechanism: An attacker or unwanted application registers a Launch Agent that points to malicious code, a renamed payload or a legitimate binary with altered behaviour, then relies on normal login activity to re-establish execution.
Impact: The result can be persistent user-session compromise, repeated credential exposure, continued malware execution and delayed detection because the process may look like ordinary desktop software.
Standards & Framework Alignment
This section maps relevant standards and security frameworks to the operational risks and controls described in this guidance.
MITRE ATT&CK addresses the attack and risk surface, while NIST SP 800-53 Rev 5 and CIS Controls v8 set the governance and control requirements practitioners need to meet.
| Framework | Control / Reference | Relevance |
|---|---|---|
| MITRE ATT&CK | T1543.001 — Launch Agent | Launch Agents are a macOS persistence technique in ATT&CK. |
| Recommendation — Map unexpected Launch Agents to T1543.001 and hunt for user-session persistence activity. | ||
| NIST SP 800-53 Rev 5 | CM-7 — Least Functionality | Launch Agents expand what runs in a user session and should be minimized. |
| SI-4 — System Monitoring | Detecting malicious Launch Agents depends on endpoint and startup monitoring. | |
| AU-6 — Audit Record Review, Analysis, and Reporting | Launch Agent investigations rely on log review and event correlation. | |
| Recommendation — Limit approved autostart entries to the minimum required user functionality. Monitor Launch Agent creation, modification and execution for suspicious persistence. Correlate launchd, file and process logs to confirm whether a Launch Agent is expected. | ||
| CIS Controls v8 | CIS-8 — Audit Log Management | Startup persistence on macOS is best detected through logged process and file activity. |
| Recommendation — Centralize and review logs that show Launch Agent creation and execution. | ||
Practitioner Guidance
What to watch for: Treat Launch Agents as inventory items, not just startup clutter. Each plist should have a clear business purpose, an expected owner and a path that matches approved software locations, especially on shared or high-value endpoints.
Governance implication: Organisations should decide who can create or modify user-level autostart entries, because that decision directly affects persistence risk. Where approval is weak, attackers and unwanted software gain an easy place to survive normal endpoint resets.
Practitioner takeaway: The most effective control is not to ban Launch Agents, but to know which ones should exist, which binaries they launch and which users are allowed to add them.
Related resources from NHI Mgmt Group
- How should security teams detect macOS spyware that uses launch agents and other persistence tricks?
- What should organisations do before allowing AI agents to write tickets or launch response actions?
- How should security teams respond when autonomous AI agents can launch supply chain attacks without a clear human operator?
- What breaks when macOS persistence is implemented through a Launch Agent that runs decoded scripts on a schedule?
Deepen Your Knowledge
Free weekly newsletter
Subscribe to the NHI & AI Identity Journal
The latest on NHI and Agentic AI security – articles, research, breaches, news and events every week.
Bonus 33% off our NHI Course when you subscribe.
Reviewed and updated by the NHIMG editorial team on November 18, 2026.
NHI Mgmt Group — the #1 independent authority on Non-Human Identity, IAM, and Agentic AI security. nhimg.org