The Agent Skills Directory

[COMMAND_EXECUTION]: The skill executes local shell commands and Node.js scripts to perform codebase analysis. It uses git for history and change detection, and utilities like grep, find, and ls to explore the file system. These operations are limited to the local project environment.
[DATA_EXFILTRATION]: The mapping process involves reading codebase configuration and source files. All extracted information is written to the local .planning/ directory. No network exfiltration or remote communication patterns were identified.
[CREDENTIALS_UNSAFE]: The skill does not contain hardcoded credentials. It is designed to document environment variable names and configuration schemas rather than extracting sensitive values.
[PROMPT_INJECTION]: The skill processes untrusted codebase data to generate summaries, creating a potential surface for indirect prompt injection. Data is ingested from project source files into the mapping agents and the scan-codebase.cjs script. While the architecture uses parallel agents to separate focus areas, and the scanner strips comments, the resulting documentation is stored in plain text for use by other agents.

kata-map-codebase