The Agent Skills Directory

[SAFE]: Extensive review of the instructions and documentation confirms the absence of malicious patterns, prompt injections, or unauthorized data access attempts. The skill operates within its defined scope of codebase hygiene and maintenance.\n- [EXTERNAL_DOWNLOADS]: The skill includes instructions for installing the fallow CLI through standard package managers such as npm and cargo. These downloads are directed to the vendor's official packages and are necessary for the skill's primary functions.\n- [COMMAND_EXECUTION]: The skill utilizes command-line tools to conduct analysis and remediation. Security is reinforced by strict operational guidelines that require the agent to provide a dry-run summary and obtain user consent before making any persistent changes to the project files.\n- [SAFE]: Although the skill processes external codebase files (untrusted data) and has code-modification capabilities, the risk of indirect prompt injection is effectively mitigated by mandatory boundary markers (dry-run previews) and human-in-the-loop verification requirements.

fallow-analysis