The Agent Skills Directory

[PROMPT_INJECTION]: The skill identifies a surface for indirect prompt injection by ingesting data from subagents and interpolating it into subsequent agent prompts.\n- Ingestion points: Reads output.yaml and plan.md from multiple subagent directories, as well as spec.md and dispatch.yaml inputs.\n- Boundary markers: Instructions differentiate between the static plan.md on disk and the dynamic prompt injection of "Upstream Context."\n- Capability inventory: The skill performs file system modifications, git operations, and execution of build/test commands via the orchestrator agent.\n- Sanitization: No specific sanitization or filtering of external/subagent content is mentioned before it is processed by downstream agents.\n- [COMMAND_EXECUTION]: The skill facilitates the execution of shell commands for building, testing, and linting the project. These commands are sourced from project configuration files or user-supplied manifests.

dispatch