The Agent Skills Directory

[EXTERNAL_DOWNLOADS]: The skill downloads a portable Apache HTTP Server distribution from https://www.apachelounge.com/download/. It dynamically parses the download page to identify and fetch the latest Win64 zip archive.
[REMOTE_CODE_EXECUTION]: After downloading and extracting the Apache archive to a local directory, the skill executes the resulting httpd.exe binary to start the web service.
[COMMAND_EXECUTION]: The provided scripts use system-level tools (Start-Process in PowerShell and subprocess.Popen in Python) to manage the background lifecycle of the Apache process and perform environment configuration.
[CREDENTIALS_UNSAFE]: The skill accepts database passwords via command-line arguments (-Password) and writes them into a local XML configuration file (default.vrd). Passwords passed as CLI arguments are potentially visible to other users or processes through system monitoring tools.
[INDIRECT_PROMPT_INJECTION]: The skill ingests configuration data from a .v8-project.json file located in the project root to resolve database and server parameters.
Ingestion points: The agent is instructed to read .v8-project.json in SKILL.md, and the supporting scripts (web-publish.ps1 and web-publish.py) parse this file to find database records.
Boundary markers: No explicit delimiters or boundary instructions are used when processing the JSON content.
Capability inventory: The skill possesses capabilities for network downloads, file system writes, and execution of arbitrary binaries (the downloaded Apache).
Sanitization: While the scripts perform basic path cleaning, there is no validation or sanitization of the values retrieved from the JSON file before they are used in commands or configuration files.

web-publish