The Agent Skills Directory

[COMMAND_EXECUTION]: The skill executes the parallel-cli tool via Bash, passing the user-provided run_id as the $ARGUMENTS variable.
Evidence: Found in SKILL.md within the bash code block: parallel-cli research poll "$ARGUMENTS" --json.
Risk: While the argument is double-quoted, this creates a surface for potential command injection if the input contains subshell execution patterns (e.g., backticks or dollar-parens) that the shell might evaluate.
[PROMPT_INJECTION]: The skill is susceptible to indirect prompt injection because it processes and displays data fetched from an external research task.
Ingestion points: The output (stdout) of the parallel-cli research poll command is ingested into the agent context.
Boundary markers: None identified; the skill instructions simply state to "Present results in a clear, organized format."
Capability inventory: The skill has access to the parallel-cli tool via Bash.
Sanitization: No sanitization or filtering of the external research content is performed before presentation to the agent.
[EXTERNAL_DOWNLOADS]: The skill references a setup procedure for the vendor's command-line interface.
Evidence: The documentation directs users to run /parallel:setup if the tool is missing.
Note: This is a standard vendor resource and setup pattern for this skill's functionality.

result