The Agent Skills Directory

[DATA_EXPOSURE]: The skill accesses project-specific directories such as _wds/tools/memory/ and {output_folder}/E-Development/ to maintain state and scan for project context. These actions are aligned with the agent's role as a build manager and do not target sensitive user credentials or system files.
[COMMAND_EXECUTION]: The skill describes a workflow involving code implementation, committing changes, and verification steps. This implies standard use of development tools (e.g., git, compilers, or test runners) within the user's project directory.
[PROMPT_INJECTION]: As a code-writing agent, this skill inherently processes external data (source code and work orders) which serves as a surface for indirect prompt injection.
Ingestion points: Codebase scanning of src/, app/, or storefront/ directories, and reading project-context.md files.
Boundary markers: No explicit delimiters or instructions to ignore embedded commands were found in the provided configuration.
Capability inventory: The agent possesses the capability to modify the filesystem, execute build commands, and perform version control operations.
Sanitization: No explicit sanitization logic for third-party code content is defined in the skill instructions.

wds-agent-mimir-builder