The Agent Skills Directory

[COMMAND_EXECUTION]: The skill is primarily designed to manage ledger artifacts by executing shell commands via the yy or yylo-ledger CLI tools. This includes creating records (artifact create), searching (artifact search), and retrieving history (artifact history). This behavior is consistent with the skill's stated purpose.
[INDIRECT_PROMPT_INJECTION]: The skill features a data ingestion surface where user-supplied evidence or artifact descriptions are processed via the $ARGUMENTS placeholder.
Ingestion points: External data enters the agent context through the $ARGUMENTS variable at the end of SKILL.md.
Boundary markers: The prompt does not utilize explicit delimiters (like XML tags or triple quotes) to separate user input from instructions.
Capability inventory: The skill has the capability to execute shell commands and read local files (e.g., via the --file flag).
Sanitization: The skill text asserts that the ledger tool itself implements security checks to reject unsafe schemes, directory traversal attempts, and known secret patterns.
[DYNAMIC_CONTEXT_INJECTION]: The SKILL.md frontmatter contains the enable-shell-directives: true configuration. This enables the agent platform to execute shell commands at load-time if specified by the !command syntax. While no such directives are currently utilized in the skill body, the capability is active in the environment metadata.

artifact-yylo