The Agent Skills Directory

[COMMAND_EXECUTION]: The skill executes the shell command parallel-cli research run "$ARGUMENTS". Because the $ARGUMENTS variable is populated directly from user input and interpolated into a bash command, an attacker could perform command injection by using shell metacharacters (such as ;, |, or $()) within their research request. This could allow for arbitrary code execution on the host system.
[PROMPT_INJECTION]: The skill is susceptible to indirect prompt injection through the processing of untrusted data from the web.
Ingestion points: Data is ingested via parallel-cli research poll, which prints an executive summary to stdout and writes detailed reports to $FILENAME.md and $FILENAME.json.
Boundary markers: The skill does not employ any boundary markers, delimiters, or explicit instructions to the agent to ignore potentially malicious directions embedded within the research results.
Capability inventory: The skill utilizes the Bash tool to execute parallel-cli commands, providing a surface for further exploitation if the agent is manipulated.
Sanitization: There is no evidence of sanitization, filtering, or validation of the external content before it is displayed to the user or stored in the file system.

parallel-deep-research