The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill takes user input via the $ARGUMENTS variable to generate feature specifications. This input is interpolated directly into the instructions, which could allow maliciously crafted input to influence the output or agent behavior.
Ingestion points: User-provided feature descriptions or arguments processed in SKILL.md.
Boundary markers: The skill does not use specific delimiters or instructions to ignore embedded commands within the input data.
Capability inventory: The agent is instructed to create directories and write files to the local file system (specs/ directory).
Sanitization: There is no evidence of filtering or sanitization of the user input before it is used.
[COMMAND_EXECUTION]: The skill workflow includes specific instructions for file system management that resemble shell commands.
Evidence: The Workflow section in SKILL.md includes steps to Create directory: specs/feature-name/ and Copy assets/spec.md to specs/feature-name/README.md.

spec-writer