The Agent Skills Directory

[PROMPT_INJECTION]: The skill possesses an indirect prompt injection surface as it processes external content and interpolates it into an image generation prompt.
Ingestion points: Untrusted data enters the context through user-provided content or file paths, which are saved to source.md in Step 1.2.
Boundary markers: The base-prompt.md template uses the {{CONTENT}} placeholder to insert source data without enclosing it in distinct delimiters or providing instructions to ignore potentially malicious embedded directives.
Capability inventory: The skill has the capability to write to the file system (creating analysis, structured content, and prompt files) and trigger external image generation tools.
Sanitization: No specific filtering, validation, or escaping of the user-provided content is performed before it is used in prompt construction.
[COMMAND_EXECUTION]: The skill uses shell commands to verify the execution environment and load user preferences.
Evidence: Step 1.1 in SKILL.md contains Bash commands using test -f to check for the existence of an EXTEND.md configuration file in both the project's local directory and the user's home directory.

baoyu-infographic