Back to skill

Security audit

Loboot

Security checks for vulnerabilities and agentic risk

Overview

This skill gives cautious instructions for managing Loboot-owned per-user services and discloses the persistence and runtime effects it can control.

Install this only if you intend to use Loboot to manage per-user native services. Review each dry-run carefully before authorizing a mutation, avoid putting secrets in arguments or manifests, and remember that enabling a service can affect future login or user-manager activation.

Vulnerability Patterns
  • Skill Instruction HijackingAlters the agent's session goals or safety constraints when the skill loads
  • Agent Memory PoisoningWrites attacker-controlled rules into memory that affect later sessions
  • Remote Payload Retrieval and ExecutionFetches external code whose behavior can change after review
  • Embedded Malicious CodeShips malicious scripts inside the skill and executes them locally
  • Unauthorized Access and Privilege EscalationObtains permissions beyond the task's legitimate needs
Vulnerability Patterns
  • Rogue AgentSelf-Modification, Session Persistence
  • Trigger AbuseOverly Broad Trigger, Shadow Command Trigger, Keyword Baiting Trigger
  • Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
  • Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
Findings (2)

Ae1

High
Category
analysis-evasion
Confidence
100% confidence
Finding

Referenced artifact was not completely inspected

Content

Scanner excerpt · SKILL.md (reported line 35)May include surrounding context.

md
Read [references/safety.md](references/safety.md) before any mutation or diagnosis involving partial progress, ownership, or native activation. Logs, Windows, s

Session Persistence

Medium
Category
Rogue Agent
Confidence
80% confidence
Finding

Skill establishes unauthorized persistence across sessions via cron jobs, startup scripts, or state files. Session persistence allows an attacker to maintain access beyond the current interaction.

Content

Scanner excerpt · references/safety.md (reported line 24)May include surrounding context.

md
## Platform and execution semantics

- Linux uses systemd `--user`; live operations require a reachable user manager, with systemd 240+ for the generated contract. A disabled service can be explicitly started. User-manager lifetime, overlapping logins, or linger can outlast logout; Loboot never modifies linger.
- macOS uses launchd in the current `gui/<uid>` graphical session. Start/restart requires enabled preference. Stop unloads with `bootout`; disable requires unloaded state. A process that finished can still have a loaded job.
- `once` means no automatic retry, not permanent deduplication or exactly one run per login. Explicit starts and later activation can run it again. Service restart policies can respawn repeatedly.
- Direct argv and TOML args are literal, but the invoking shell can expand unquoted input and a program can interpret its own arguments. Do not add shell evaluation without an explicit need and authorization.
- Every path component must be non-symlinked. Use physical paths such as macOS `/private/tmp` instead of `/tmp`; do not replace the user's unrelated symlinks. Linux also rejects quotes, backslashes, and glob characters in the executable pathname.