The Agent Skills Directory

[COMMAND_EXECUTION]: The skill generates shell commands to create symbolic links and modify the .gitignore file. These actions are core to the skill's functionality and include safety checks such as verifying path existence and requiring user confirmation before proceeding.
[REMOTE_CODE_EXECUTION]: The skill executes git clone using URLs provided by the user. The AskUserQuestion tool is used to ensure the user confirms the clone location and the source repository, which mitigates the risk of unauthorized remote code downloads.
[PROMPT_INJECTION]: A surface for indirect prompt injection exists when the skill reads the wiki-manifest.yaml file from a newly linked or cloned repository. Ingestion points: .wiki/wiki-manifest.yaml (referenced in SKILL.md). Boundary markers: None specified for the content of the manifest file. Capability inventory: git clone, symbolic link creation via shell commands, and file modification for .gitignore. Sanitization: No explicit sanitization or instruction filtering is mentioned for the contents of the manifest file.

wiki-link