The Agent Skills Directory

[SAFE]: A comprehensive analysis of the skill's instructions and configuration found no evidence of malicious intent or safety bypass attempts.
[COMMAND_EXECUTION]: The skill utilizes the squirrel CLI to perform its core functions. It defines a structured set of commands including audit, report, and init to manage the auditing process.
[EXTERNAL_DOWNLOADS]: The skill documentation references https://squirrelscan.com for CLI installation and rule documentation. This is the official domain of the tool provider and is documented as a prerequisite for using the skill.
[PROMPT_INJECTION]: The skill possesses an indirect prompt injection surface because it processes data from external websites to suggest local file modifications.
Ingestion points: Audit data is ingested from the output of the squirrel report command (SKILL.md).
Boundary markers: The instructions do not define explicit boundary markers or delimiters for the tool output.
Capability inventory: The skill has access to the Read, Edit, Grep, Glob, and Bash tools, which allow for local file modification and command execution.
Sanitization: There are no instructions for sanitizing or validating the audit reports before they are processed by the agent.
Mitigation: The skill instructions explicitly mandate that the agent must "ask the user to confirm before making changes" and "Pause for judgment" for ambiguous issues, which serves as a critical safety checkpoint.

audit-website