The Agent Skills Directory

Indirect Prompt Injection Surface: The skill analyzes external content from the local project environment to determine integration requirements. This represents a surface where instructions embedded in data could influence the agent.
Ingestion points: The 'Explore agent' in Phase 2.1 reads project source code files (src/**/*.{ts,tsx,js,jsx,vue,astro}) and the .datamodel-manifest.json file. Phase 6.2 also allows for the ingestion of user-provided diagrams or text descriptions.
Boundary markers: The instructions provided to the sub-agents do not explicitly define boundary markers or include directives to ignore natural language instructions that might be present within the analyzed code or data files.
Capability inventory: The skill possesses significant capabilities, including the ability to write and edit files (Write, Edit), execute build scripts via the shell (Bash), and commit changes to version control (git commit).
Sanitization: There is no evidence of sanitization or filtering of the content analyzed from the source files before it is processed by the integration agents.
Automated Command Execution: The skill performs several automated shell operations as part of its standard workflow.
Build Execution: Phase 5.2 executes npm run build to verify the generated integration code. This executes scripts defined in the project's package.json.
Version Control: Phases 4.4 and 6.5 automatically stage and commit changes using git commit. While standard for development tools, users should ensure the changes are reviewed before final deployment.

integrate-webapi