The Agent Skills Directory

[COMMAND_EXECUTION]: The skill relies on executing shell commands through the kd CLI tool to perform project initialization, component creation, and deployment. These actions are documented and essential to the skill's workflow.
[INDIRECT_PROMPT_INJECTION]: The skill has an attack surface for indirect prompt injection as it ingests untrusted data from project files and API responses.
Ingestion points: Project metadata files (.page-meta.kwp, .js-meta.kwc, .kws), environment configuration (~/.kd/config.json in scripts/_shared.mjs), and responses from external API endpoints.
Boundary markers: The instructions do not define specific delimiters to isolate processed data from the agent's instructions.
Capability inventory: The skill can execute CLI commands (via SKILL.md), perform file system operations, and make network requests (via scripts/menu-api.mjs and scripts/meta-query-api.mjs).
Sanitization: No specialized sanitization or filtering is applied to data retrieved from external sources before it is used to influence the agent's behavior.
[EXTERNAL_DOWNLOADS]: The scripts provided (menu-api.mjs, meta-query-api.mjs, _shared.mjs) make network requests to Kingdee's official domains to manage authentication and retrieve project metadata. These operations are consistent with the vendor's intended functionality.

kwc-project-scaffold