The Agent Skills Directory

[EXTERNAL_DOWNLOADS]: The skill instructions direct the agent to clone a repository from GitHub (https://github.com/legalize-dev/legalize-es.git). This fetches a large volume of external Markdown files and their complete Git history into the execution environment.
[COMMAND_EXECUTION]: The skill provides numerous examples of shell commands that the agent is expected to run, including git, grep, sed, cat, and ls. These commands are used to interact with the filesystem and the Git database of legislation.
[DYNAMIC_EXECUTION]: The provided Python scripts use subprocess.run to dynamically execute git commands (such as git log, git diff, and git show). Arguments for these commands, such as file paths and commit hashes, are constructed at runtime from script variables and potentially from the content of the legislation files themselves.
[INDIRECT_PROMPT_INJECTION]: The skill is designed to ingest and process a large number of Markdown files from an external repository, creating a vulnerability to indirect prompt injection.
Ingestion points: Thousands of .md files located in the spain/ directory of the cloned repository.
Boundary markers: No specific delimiters or "ignore" instructions are used when parsing the file content.
Capability inventory: The skill utilizes shell command execution via subprocess.run and standard file system operations (reading and writing files).
Sanitization: There is no evidence of sanitization or filtering to prevent the agent from following instructions potentially embedded within the legislative text or metadata.

legalize-es-spanish-legislation