The Agent Skills Directory

[COMMAND_EXECUTION]: The skill performs several shell commands to reset and clean the target workspace, including git clean -fd, git checkout origin/main -- ., and rm -rf node_modules package-lock.json bun.lock. While documented as part of the migration workflow, these are destructive operations.
[DYNAMIC_EXECUTION]: The skill executes a local TypeScript script using tsx (packages/blocks-cli/scripts/migrate.ts). This allows for the execution of code contained within the repository against a target directory.
[INDIRECT_PROMPT_INJECTION]: The skill is susceptible to indirect prompt injection as it processes external source code from a target site workspace.
Ingestion points: The agent reads and analyzes the source code of a user-specified target workspace (referenced in SKILL.md).
Boundary markers: None are specified for the processing of external code or the generation of the MIGRATION_REPORT.md.
Capability inventory: The skill has capabilities to modify the file system (git, rm), install packages (bun install), and execute scripts via npx and tsx (referenced in SKILL.md).
Sanitization: There is no mention of sanitizing the content of the target site before it is processed by the migration script or reviewed by the agent.

run-migration