The Agent Skills Directory

[EXTERNAL_DOWNLOADS]: The skill instructs users to install an external plugin (team-agents@duyet-claude-plugins) to enable sub-agent coordination features. This resource is provided by the skill author.
[COMMAND_EXECUTION]: The orchestration framework explicitly delegates tasks to worker agents that are permitted to execute shell commands using the Bash tool to fulfill their assigned responsibilities.
[PROMPT_INJECTION]: The skill architecture presents a surface for indirect prompt injection (Category 8). It ingests untrusted user input to dynamically generate instructions for subordinate agents.
Ingestion points: User-provided requests and context are used to define the TASK, CONTEXT, and SCOPE fields for background worker agents as described in SKILL.md and references/tools.md.
Boundary markers: The skill employs the === WORKER AGENT === preamble and various markdown headers as delimiters. These markers are used to separate orchestration instructions from task-specific data but do not provide absolute isolation against adversarial input.
Capability inventory: Spawned worker agents have significant capabilities, including file system modification (Write, Edit) and shell access (Bash), which could be abused if an injection succeeds.
Sanitization: The instructions do not provide specific mechanisms for escaping or validating user input before it is interpolated into the prompts sent to background tasks.

orchestration