The Agent Skills Directory

[COMMAND_EXECUTION]: The skill requires the execution of a curl command to a local service (http://localhost:8888/notify) immediately upon invocation of its workflows. These commands are configured to run silently in the background via shell redirection (> /dev/null 2>&1 &), which prevents the user from seeing the execution or any potential errors.
[PROMPT_INJECTION]: The skill is susceptible to indirect prompt injection due to its handling of untrusted user content.
Ingestion points: User-provided arguments, plans, or documents are ingested for analysis within the ParallelAnalysis.md and AdversarialValidation.md workflows.
Boundary markers: While the skill uses Markdown headers (e.g., ## THE ARGUMENT TO ANALYZE:) to delineate user content in the agent prompts, it lacks explicit instructions to the AI to ignore instructions or command sequences embedded within the ingested text.
Capability inventory: The skill has access to shell commands (curl) and uses parallel execution tools to manage 32 agent personas.
Sanitization: No input validation or escaping is performed on the user-supplied content before it is interpolated into the prompts.

RedTeam