The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill builds a behavior inventory by reading project files like README, docs, and source code. Maliciously crafted content in these files could theoretically influence the agent's logic during the analysis or test generation phases.
Ingestion points: Project files including README, docs/, CHANGELOG, and source code.
Boundary markers: None explicitly provided to separate ingested data from instructions in the prompt.
Capability inventory: File writing (CRITERIA.md), package installation (npm install), and shell command execution (npx playwright test).
Sanitization: No explicit sanitization or filtering of ingested file content is mentioned.
[COMMAND_EXECUTION]: Executes shell commands to configure the testing environment and run the test suite.
Evidence: Use of npm i -D @playwright/test, npx playwright install --with-deps chromium, and npx playwright test.
[DYNAMIC_EXECUTION]: Generates and runs TypeScript/JavaScript test files based on user-agreed criteria derived from codebase analysis.
Evidence: The "Write" phase creates test files that are subsequently executed during the "Run" phase.
[EXTERNAL_DOWNLOADS]: Installs the Playwright framework and browser binaries from official package registries.
Evidence: Downloads @playwright/test and chromium dependencies via npm and npx.

acceptance-test