The Agent Skills Directory

[COMMAND_EXECUTION]: The skill uses the bash tool to perform git operations and execute a local state management script located at ~/.claude-octopus/plugin/scripts/state-manager.sh.- [INDIRECT_PROMPT_INJECTION]: The skill processes user-supplied task descriptions and embeds them into system outputs, creating a vulnerability surface.
Ingestion points: User-provided task descriptions from the /octo:quick command or skill invocation as seen in the usage examples.
Boundary markers: There are no explicit instructions for the agent to use boundary markers or delimiters when handling the user's task input.
Capability inventory: The skill utilizes bash, Edit, and Write tools, which grant the agent significant control over the local environment and file system.
Sanitization: The implementation steps do not specify sanitization or escaping of the $TASK_DESCRIPTION variable, which is interpolated directly into a shell heredoc during summary generation (cat > summary.md <<EOF ... $TASK_DESCRIPTION ... EOF).

octopus-quick