The Agent Skills Directory

[COMMAND_EXECUTION]: The skill instructions and examples direct the agent to execute shell commands such as npm run build, node --test, and git status to verify task completion and manage the repository state.
[INDIRECT_PROMPT_INJECTION]: The skill defines a workflow that ingests untrusted data from user requests and repository files. This content is interpolated into agent instructions and used to determine which tools to run or tasks to delegate.
Ingestion points: Processes user-provided task descriptions and repository file content as described in the SKILL.md workflow.
Boundary markers: The skill uses instructional tags like <Purpose> and <Execution_Policy> for its own structure but does not specify explicit delimiters or isolation for untrusted data processed at runtime.
Capability inventory: The skill possesses significant capabilities including shell command execution (npm, node, git), file system access, state management via rcs_state tools, and multi-agent delegation.
Sanitization: No explicit input validation or sanitization mechanisms are defined for external content.

ultrawork