The Agent Skills Directory

[COMMAND_EXECUTION]: The skill executes npx codebase-intelligence with various flags to extract project structure, hotspots, and dependency metrics for software analysis.
[EXTERNAL_DOWNLOADS]: The skill fetches documentation from the author's GitHub repository at https://raw.githubusercontent.com/bntvllnt/codebase-intelligence/main/llms.txt and invokes external packages via npx at runtime.
[INDIRECT_PROMPT_INJECTION]: The skill is designed to process untrusted user-provided content and file data. \n
Ingestion points: Target file paths and topic descriptions provided in the analysis command (SKILL.md). \n
Boundary markers: The prompt template uses a TARGET: [content] label but lacks robust delimiters or explicit instructions to ignore embedded instructions. \n
Capability inventory: The skill utilizes file system read access and the ability to execute subprocesses via npx. \n
Sanitization: No explicit input sanitization is described, although agents are instructed to specifically identify misuse and adversarial patterns.
[DYNAMIC_EXECUTION]: The skill dynamically generates and executes commands using the codebase-intelligence package to produce structured metadata about software projects at runtime.

analyze