The Agent Skills Directory

[COMMAND_EXECUTION]: The skill executes powershell.exe to run a script located at .claude/skills/epf-build/scripts/epf-build.ps1. It uses dynamic arguments derived from local configuration and user input.
[CREDENTIALS_UNSAFE]: The skill manages database authentication by accepting -UserName and -Password parameters. The provided examples show passwords being passed as plaintext command-line arguments, which can result in exposure through system process monitoring tools or shell history.
[PROMPT_INJECTION]: The skill is susceptible to indirect prompt injection via the .v8-project.json file.
Ingestion points: Reads configuration values such as v8path and database connection parameters from .v8-project.json in the project root.
Boundary markers: No markers or instructions are present to prevent the agent from interpreting malicious content within the JSON file as commands.
Capability inventory: The skill has the capability to execute PowerShell commands, search the filesystem for executables, and interact with 1C databases.
Sanitization: There is no evidence of validation or sanitization for the paths or parameters extracted from the configuration file before they are used in command construction.

erf-build