The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill is vulnerable to indirect prompt injection because it processes untrusted input data—such as file paths, registry keys, and user accounts—and interpolates them directly into shell command templates.
Ingestion points: The MalwareRemediation.remove_malware method in scripts/remediation_utils.py accepts a dictionary of malware_artifacts containing lists of files, registry keys, and processes. The AccessRemediation.full_credential_reset method accepts a list of users.
Capability inventory: The skill generates powerful system management commands across multiple files, including taskkill, pkill, sc delete, reg delete, rm, net user, and userdel.
Sanitization: Input strings are interpolated into commands (e.g., f"rm -f \"{f}\"") without shell escaping or sanitization, allowing for potential command injection if the input contains subshell syntax like $(...) or backticks.
Boundary markers: The generated commands do not include boundary markers or instructions to the executing environment to treat the interpolated values as data only.
[COMMAND_EXECUTION]: The skill defines a large surface for high-privilege command execution by generating remediation scripts for Windows and Linux systems. While the documentation in SKILL.md states that it "Generates commands/procedures, does not execute directly," the primary intent is for an agent or administrator to run these commands, which could lead to significant system changes or unintended consequences if misused.

remediation