The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill's primary workflow involves ingesting and analyzing untrusted data from project files located in /mnt/user-data/uploads. This creates an attack surface for indirect prompt injection, where malicious instructions hidden within code comments, configuration metadata, or project documentation could influence the agent's logic or behavior during analysis. \n
Ingestion points: The skill reads file contents from /mnt/user-data/uploads and specifically targets configuration files like package.json, pubspec.yaml, and requirements.txt to detect frameworks and conventions. \n
Boundary markers: The instructions do not define explicit delimiters or 'ignore' directives to prevent the agent from accidentally obeying instructions embedded within the codebase it is analyzing. \n
Capability inventory: The skill uses discovery commands such as ls and view to inspect the project filesystem and retrieve file contents. \n
Sanitization: There is no mention of filtering, escaping, or validating the ingested content before it is processed by the agent to build the enhanced requirements.

prompt-enhancer