The Agent Skills Directory

[COMMAND_EXECUTION]: The skill enables and documents the execution of external commands through the Bash tool (used for Roswell, Qlot, and SBCL) and the sb-ext:run-program interface. These capabilities are expected for managing Lisp implementations and build workflows.
[DYNAMIC_EXECUTION]: The skill describes advanced dynamic features of Common Lisp, such as macros, Foreign Function Interfaces (FFI) via sb-alien and cffi, and read-time code evaluation using the #. reader macro. It provides specific warnings about load-order complexities associated with read-time evaluation.
[INDIRECT_PROMPT_INJECTION]: The skill processes project-specific source files and system definitions, which presents an attack surface for indirect prompt injection if the files are untrusted.
Ingestion points: Analyzes ASDF system definitions and Lisp source files using the Read tool.
Boundary markers: No specific delimiters or "ignore instructions" markers are defined for the code the agent reads.
Capability inventory: The agent has the ability to execute shell commands, edit files, and invoke foreign code.
Sanitization: While the skill provides security best practices for developers (e.g., preventing format string vulnerabilities), it does not document internal sanitization logic for the data it ingests.
[EXTERNAL_DOWNLOADS]: The skill references standard package management workflows using qlot and roswell. These tools interact with well-known community repositories like Quicklisp and Ultralisp to manage dependencies.

common-lisp-ecosystem