The Agent Skills Directory

[COMMAND_EXECUTION]: The skill instructs the agent/user to perform troubleshooting by executing xmake commands with diagnostic flags (e.g., xmake -vD, xmake f -c). It also includes commands for cleaning the project environment, such as xmake clean --all and rm -rf .xmake to remove local cache directories.
[DYNAMIC_EXECUTION]: The xmake l cli.bisect functionality is detailed, which involves the dynamic execution of shell commands or Lua scripts in a loop to identify the first bad commit in a regression. The skill provides examples of Lua scripts using os.exec and os.iorun for automated testing during the bisect process.
[EXTERNAL_DOWNLOADS]: The skill mentions the xrepo package manager, which downloads dependencies and tools, specifically referencing the installation and setup of the EmmyLua debugger environment (xrepo env -b emmylua_debugger).
[INDIRECT_PROMPT_INJECTION]: The skill documents an attack surface where the cli.bisect tool processes user-provided shell commands and scripts.
Ingestion points: Shell commands passed via the -c flag and Lua scripts via the -s flag in the xmake l cli.bisect utility.
Boundary markers: None are present in the documentation; instructions are processed directly by the tool's execution engine.
Capability inventory: The skill uses tools capable of subprocess execution (os.exec), filesystem modification (rm -rf), and network access (via xrepo).
Sanitization: No input sanitization or validation is mentioned for the command strings or script paths provided to the bisect tool.

xmake-troubleshooting