The Agent Skills Directory

[PROMPT_INJECTION]: The skill processes untrusted content from WebSearch snippets and WebFetch results from various company websites, creating a vulnerability surface for indirect prompt injection. * Ingestion points: Data enters the agent's context through web search results and fetched website text as described in the browser scraping guide. * Boundary markers: There are no explicit instructions or delimiters defined to isolate external content or prevent the agent from following instructions found within it. * Capability inventory: The skill has the capability to write files and execute shell commands for directory management. * Sanitization: The skill does not implement validation or escaping of the web-sourced data before processing it for scoring or inclusion in reports.
[COMMAND_EXECUTION]: The skill utilizes shell commands to manage the local file system and generate output. * Evidence: Instructions in the main skill file specify the use of mkdir -p to create output directories. * Evidence: The export templates documentation suggests the use of bash heredocs to create CSV and Markdown files, which involves executing code with data derived from external sources.

lead-research-assistant