The Agent Skills Directory

[COMMAND_EXECUTION]: The skill executes a shell script located at ~/scripts/create_worktree.sh. This external dependency is used to create new development environments based on provided branch names.
[COMMAND_EXECUTION]: The skill utilizes the humanlayer CLI tool to launch automated agent sessions (using the 'opus' model) for code implementation.
[INDIRECT_PROMPT_INJECTION]: The skill is susceptible to indirect injection because it interpolates potentially untrusted data (such as branch names or file paths) directly into a command string for the humanlayer agent.
Ingestion points: Variable inputs like BRANCH_NAME and $FILEPATH are embedded into instructions for the next agent session.
Boundary markers: The instructions lack explicit delimiters or "ignore" directives to prevent the agent from misinterpreting data within the file path as part of the command.
Capability inventory: The skill has the ability to execute shell scripts and launch autonomous agent sessions.
Sanitization: There is no evidence of validation or sanitization for the branch names or file paths before they are executed in the shell or passed to the agent.

worktree