The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The scripts/bundle_analyzer.py utility parses local project files such as package.json and source code to identify optimization opportunities. This creates a standard attack surface for indirect prompt injection if the analyzed files were to contain malicious instructions targeting the AI agent's interpretation of the scan results.
Ingestion points: scripts/bundle_analyzer.py reads and processes the content of package.json and various source files (.ts, .tsx, .js, .jsx).
Boundary markers: The script output lacks specific boundary markers or 'ignore' instructions to separate analyzed file data from the tool's generated report.
Capability inventory: The skill possesses file-writing capabilities through its project scaffolding and component generation scripts, though these are limited to the intended local development directories.
Sanitization: The tool uses standard JSON parsing and regular expressions for analysis without specialized sanitization for instruction-bypass strings.

senior-frontend