The Agent Skills Directory

[COMMAND_EXECUTION]: The scripts/install.sh script executes shell commands to install the pyyaml Python package and the skill itself via npx. It also uses a heredoc to execute a Python-based bundle validator.
[EXTERNAL_DOWNLOADS]: The skill downloads the pyyaml library from official Python package registries during setup. It also fetches its own code and configuration from the author's GitHub repository (akillness/jeo-skills) using the npx skills add command.
[DYNAMIC_EXECUTION]: The skill includes Python scripts embedded within the documentation and the installer that are executed at runtime to validate file structures and parse YAML frontmatter. It correctly utilizes yaml.safe_load() for secure parsing of data.
[INDIRECT_PROMPT_INJECTION]: The skill's primary utility involves ingesting and processing external Markdown files ('knowledge atoms'). This creates an attack surface where instructions embedded in these files could influence the agent's behavior.
Ingestion points: The skill provides Python logic in SKILL.md and references/usage.md to recursively read and process Markdown files from a directory.
Capability inventory: The skill is configured with access to Bash, python3, and WebFetch tools.
Boundary markers: There are no specific instructions or delimiters provided to warn the agent against executing commands found within the processed knowledge atoms.
Sanitization: The skill uses yaml.safe_load() to process metadata, which prevents the execution of arbitrary code through YAML tags.

okf