The Agent Skills Directory

[DYNAMIC_EXECUTION]: The skill utilizes the eval command to execute shell output from the wt session set utility. This is used to dynamically configure the environment for coordinator sessions.\n
Evidence: The instruction eval "$(wt session set coord-<work-slug>)" in SKILL.md captures and executes shell exports.\n- [INDIRECT_PROMPT_INJECTION]: The workflow relies on processing external inputs such as TaskDocument files and Agent Completion Reports. These documents could potentially contain instructions that influence agent behavior.\n
Ingestion points: The skill reads task definitions from .wt/execution/tasks and monitors worker reports via wt task report and wt msg watch.\n
Capability inventory: The agent has the ability to execute shell commands (wt, git, cargo, cmux), write to the filesystem, and manage worktrees.\n
Sanitization: The instructions mitigate this risk by requiring the agent to perform direct inspections of file contents, diffs, and build outputs rather than relying solely on the reports.\n
Boundary markers: There are no explicit delimiters or safety instructions provided for the contents of the task files.\n- [COMMAND_EXECUTION]: The skill is designed around the execution of various CLI tools to manage the development lifecycle, including repository management, build tools, and custom task runners.\n
Evidence: Extensive use of wt, git, cargo, and cmux commands across all sections of SKILL.md.

work