The Agent Skills Directory

[COMMAND_EXECUTION]: The script scripts/ask_fable.mjs uses spawnSync to execute the claude command-line utility.- [INDIRECT_PROMPT_INJECTION]: The skill accepts user-provided text from standard input and passes it to a model execution process that has access to system tools.
Ingestion points: User input is read via standard input and inherited by the spawned claude process.
Boundary markers: The skill does not apply explicit delimiters or instructions to ignore embedded commands within the user input.
Capability inventory: The model is invoked with tools such as Read, Grep, and Glob enabled through the --tools flag.
Sanitization: The skill performs no sanitization or validation of the input request before processing.- [DATA_EXPOSURE]: The script scripts/ask_fable.mjs reads a system prompt from a file located at ../../../claude-agents/fable-advisor.md, which is a relative path that accesses files outside the skill's root directory.

fable-advisor