The Agent Skills Directory

[COMMAND_EXECUTION]: The scripts scripts/executor.py and scripts/doc_generator.py utilize subprocess.run(shell=True) to execute system commands. This includes executing installation plans and running software verification checks.\n- [COMMAND_EXECUTION]: A command injection vulnerability exists in scripts/doc_generator.py and scripts/plan_generator.py because the software name (target) is interpolated directly into shell command strings (e.g., f"{target} --version") without sanitization or escaping.\n- [EXTERNAL_DOWNLOADS]: The skill is designed to fetch and install software from remote repositories using standard package managers such as apt, brew, dnf, yum, pacman, winget, and choco.\n- [PROMPT_INJECTION]: The skill exhibits a vulnerability surface for indirect prompt injection.\n
Ingestion points: Untrusted data enters the system through the software name (--target) and dependencies (--deps) parameters in scripts/plan_generator.py.\n
Boundary markers: None. There are no delimiters or instructions to treat these inputs as data rather than executable commands.\n
Capability inventory: The skill can execute arbitrary shell commands and acquire administrative privileges via sudo.\n
Sanitization: No escaping, validation, or filtering is applied to the input strings before they are incorporated into shell commands.\n- [DATA_EXFILTRATION]: scripts/env_explorer.py performs extensive system discovery, collecting sensitive environment metadata including OS details, user permissions (sudo/admin status), and directory structures. While necessary for the skill's function, this increases the risk if the data is mishandled.

install-script-generator