The Agent Skills Directory

[COMMAND_EXECUTION]: The skill uses subprocess execution to manage local development servers for testing dynamic applications.
Evidence: The SKILL.md file provides code examples using subprocess.Popen to execute commands like npm run dev.
Context: The skill enforces strict security rules, specifically the 'Iron Law' of using shell=False and array arguments to mitigate command injection risks.
[PROMPT_INJECTION]: The skill's primary function involves processing external data from web pages, creating a surface for indirect prompt injection.
Ingestion points: Untrusted data enters the agent context through page content inspection, title reading, and browser console log capture as described in SKILL.md.
Boundary markers: The skill does not define explicit delimiters or instructions to ignore embedded commands within the content retrieved from the browser.
Capability inventory: The agent possesses the Bash, Write, and Playwright automation capabilities, which could be misused if the agent inadvertently follows instructions embedded in a tested web page.
Sanitization: There is no evidence of sanitization or filtering of the extracted DOM content or console logs before the agent processes them.

webapp-testing