The Agent Skills Directory

[REMOTE_CODE_EXECUTION]: The skill facilitates the execution of arbitrary code from untrusted project files. The helper script agent-skill-ent.el uses the Emacs load function on .ent.el files discovered in the project directory, which can contain and execute any Emacs Lisp command.\n- [DYNAMIC_EXECUTION]: The functions agent-skill-ent--load-project and agent-skill-ent-tasks perform dynamic loading of executable content from computed file paths (initfile) without validation of the file's integrity or source.\n- [INDIRECT_PROMPT_INJECTION]: The skill presents an attack surface for indirect prompt injection because it is designed to ingest and act upon data from untrusted project files (.ent.el).\n
Ingestion points: Reads and executes .ent.el files found in the current project or parent directories.\n
Boundary markers: None; the file is loaded directly into the Emacs environment.\n
Capability inventory: The skill can execute shell commands via emacsclient, kill processes, and execute arbitrary Emacs Lisp via the loaded files.\n
Sanitization: No sanitization or verification of the configuration file content is performed before execution.\n- [COMMAND_EXECUTION]: The skill uses emacsclient --eval to run complex logic in a shell environment, which includes loading local scripts and executing build tasks that typically involve arbitrary shell command execution.

run-ent-task