The Agent Skills Directory

[COMMAND_EXECUTION]: The skill extensively utilizes the Bash tool to automate project tasks. This involves executing multiple internal scripts for session management (catalyst-session.sh), environment verification (check-project-setup.sh), and automated pull request management (draft-pr.sh). These scripts are located within the ${CLAUDE_PLUGIN_ROOT}/scripts/ directory, which is local to the agent's plugin environment.
[PROMPT_INJECTION]: The skill has an attack surface for indirect prompt injection because it reads and processes technical implementation plans from thoughts/shared/plans/. These plans contain phase-by-step instructions that the agent is directed to follow.
Ingestion points: Technical plan files stored in thoughts/shared/plans/ are read and processed in Step 3.
Boundary markers: None identified; the agent is instructed to follow the plan's intent and checkboxes authoritativeley.
Capability inventory: Significant access to system tools including Bash, file system writing/editing, and the ability to spawn specialized subagents for review tasks.
Sanitization: No content validation or sanitization of the input plan files is performed before the agent acts on them.

implement-plan