The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill processes external project specifications and task lists to drive its orchestration logic.
Ingestion points: Reads content from files in project-specs/ and project-tasks/ using cat and grep in SKILL.md.
Boundary markers: No explicit delimiters or instructions are provided to the agent to treat external content as untrusted data.
Capability inventory: The orchestrator can list and read files via shell commands and initiate the spawning of multiple other specialized agents with specific instructions.
Sanitization: Content from the specification files is passed directly into the context of spawned agents without escaping or validation.
[COMMAND_EXECUTION]: Uses standard shell commands in SKILL.md to manage workflow state.
Evidence: Uses ls -la, cat, and grep to verify the presence of specification files and task lists, and to count pending tasks. These commands are used solely for project state management within the local environment.

Agents Orchestrator