The Agent Skills Directory

[PROMPT_INJECTION]: The skill ingests untrusted data from the local repository (specifically docs/custom-sidebars.md and source files in Packages/) using sed and rg to inform the AI's sidebar generation logic. This creates a surface for indirect prompt injection if those repository files were to contain malicious instructions designed to hijack the agent's behavior.
Ingestion points: repo/docs/custom-sidebars.md, repo/Packages/CmuxSwiftRender/Sources/CmuxSwiftRender/SwiftViewInterpreter.swift and other source files.
Boundary markers: Absent; instructions do not specify delimiters to separate repository content from agent instructions.
Capability inventory: The skill can write files to ~/.config/cmux/sidebars/ and execute cmux CLI commands for validation and reloading.
Sanitization: No sanitization or filtering of the ingested file content is performed before processing.
[COMMAND_EXECUTION]: The skill instructions direct the agent to use several shell commands including mkdir, sed, rg, and a custom cmux CLI. These commands are used for legitimate project discovery and configuration management within the scope of the user's workspace and the ~/.config directory.

cmux-sidebar-builder