The Agent Skills Directory

[COMMAND_EXECUTION]: The script scripts/emacsctl.py utilizes subprocess.run to execute the emacsclient binary for runtime communication with the Emacs server. This is a standard and necessary function for the skill's operation.\n- [EXTERNAL_DOWNLOADS]: Documentation in SKILL.md suggests using nix shell to obtain the info documentation system in environments where it is unavailable. This leverages the well-known Nix package manager to provide legitimate system utilities.\n- [PROMPT_INJECTION]: The skill facilitates reading state and content directly from a running Emacs instance, creating a surface for indirect prompt injection attacks.\n
Ingestion points: Untrusted data enters the agent context via scripts/emacsctl.py (specifically cmd_buffer and cmd_eval functions) and query patterns defined in references/common-queries.md.\n
Boundary markers: The skill does not implement delimiters or explicit 'ignore embedded instructions' warnings when retrieving data from the Emacs runtime.\n
Capability inventory: The skill can execute arbitrary Elisp code (which can spawn shell commands) and perform file system writes via standard agent capabilities as described in the workflow instructions.\n
Sanitization: Content fetched from Emacs buffers and variables is not sanitized or validated before being presented to the agent for processing.

emacs