The Agent Skills Directory

[EXTERNAL_DOWNLOADS]: The skill identifies and automatically downloads images from arbitrary URLs extracted from Markdown documents.
[EXTERNAL_DOWNLOADS]: The scripts/download_image.py script disables SSL certificate verification when performing requests (verify=False). This allows for potential Man-in-the-Middle (MitM) attacks, which could lead to the interception or modification of downloaded assets.
[COMMAND_EXECUTION]: The skill instructs the agent to execute a local Python script with arguments (URLs) derived directly from untrusted external documents.
[PROMPT_INJECTION]: The skill is vulnerable to indirect prompt injection because it ingests and processes untrusted data from Markdown files and images (via OCR). Malicious instructions embedded in these sources could attempt to influence the agent's behavior during information extraction or reporting.
Ingestion points: External Markdown files and local/remote images.
Boundary markers: The skill lacks explicit delimiters or instructions to ignore commands potentially hidden within processed text or image content.
Capability inventory: Includes file system scanning, network downloads, and local script execution.
Sanitization: There is no evidence of validation or sanitization of data extracted from external sources before it is used in the final generated report.

recruitment-processor