The Agent Skills Directory

[REMOTE_CODE_EXECUTION]: The skill instructs the user to execute a remote shell script directly from an external URL with root privileges.
Evidence: sudo /bin/sh -c "$(curl -fsSL https://static.palera.in/scripts/install.sh)" targeting static.palera.in.
[PERSISTENCE]: The skill attempts to maintain configuration and path access across sessions by modifying shell profile files.
Evidence: echo "export THEOS=~/theos" >> ~/.zprofile && echo "export PATH=$THEOS/bin:$PATH" >> ~/.zprofile.
[PRIVILEGE_ESCALATION]: The skill makes extensive use of sudo to perform tasks that acquire administrative control over the host system.
Evidence: Using sudo for piped shell execution, moving downloaded binaries to /usr/bin, and stopping/starting system services like usbmuxd.
[EXTERNAL_DOWNLOADS]: The skill downloads executable binaries and toolchains from external repositories that are not included in the trusted vendor list.
Evidence: Downloading palera1n-linux-x86_64 from GitHub releases and moving it to a system binary path (/usr/bin/palera1n), as well as using git clone for the Theos toolchain.
[COMMAND_EXECUTION]: The skill utilizes several package managers and system tools to install unversioned software.
Evidence: Usage of pip install frida-tools, brew install, and apt install for various reverse engineering utilities.
[INDIRECT_PROMPT_INJECTION]: The skill presents a surface for indirect injection by ingesting and processing untrusted data from third-party iOS applications.
Ingestion points: App binaries, Mach-O headers, and memory strings are processed via strings, otool, dpkg-deb, and Frida's readCString() method in SKILL.md.
Boundary markers: No boundary markers or "ignore instructions" warnings are present when processing external binary data.
Capability inventory: The skill has access to powerful capabilities including root-level command execution, file system modification, and network operations (curl, ssh, iproxy).
Sanitization: There is no evidence of sanitization or validation of the data read from external applications before it is logged or processed by the agent environment.

re-ios-jb