The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill processes untrusted user-uploaded data and project files to generate requirements, which presents a surface for indirect prompt injection attacks.
Ingestion points: The skill reads files from /mnt/user-data/uploads and various project configuration files (e.g., package.json, pubspec.yaml, requirements.txt) to gather context.
Boundary markers: The skill lacks explicit instructions to ignore or delimit potentially malicious instructions embedded within the analyzed project files.
Capability inventory: The skill identifies project structures and generates text-based requirement documents. It does not perform automated code execution or network operations itself, which limits the immediate impact.
Sanitization: There is no evidence of sanitization or filtering of the content extracted from the user's project files before it is interpolated into the generated requirement document.

prompt-enhancer