Agent-native E2E runtime with verifiable safety. 13 MCP tools including alethia_propose_tests (agent generates tests from a URL), alethia_assert_safety (proves destructive actions are blocked), and the expect block: NLP primitive unique to Alethia. Zero-IPC, ~45x faster than Playwright, signed evidence packs. Works with Claude Code, Cursor, Cline.
95
94%
Does it follow best practices?
Impact
97%
2.77xAverage score across 5 eval scenarios
Advisory
Suggest reviewing before use
{
"context": "Tests whether the agent produces the correct package name (@vitronai/alethia), correct install command (npm install -g), and correct MCP server configuration (command: alethia-mcp under mcpServers.alethia). Also checks that the sample test uses correct NLP phrasing and file:// navigation.",
"type": "weighted_checklist",
"checklist": [
{
"name": "Correct npm package name",
"description": "setup_guide.md and/or mcp_config.json use the exact package name '@vitronai/alethia' (not 'alethia', 'vitronai/alethia', or any other variant)",
"max_score": 12
},
{
"name": "Global npm install command",
"description": "setup_guide.md shows the install command as 'npm install -g @vitronai/alethia'",
"max_score": 10
},
{
"name": "mcpServers key in config",
"description": "mcp_config.json contains a 'mcpServers' key at the top level",
"max_score": 8
},
{
"name": "alethia server key",
"description": "mcp_config.json has 'alethia' as the server name under mcpServers",
"max_score": 8
},
{
"name": "alethia-mcp command",
"description": "mcp_config.json sets command to 'alethia-mcp' (not 'npx alethia', 'node alethia', or another value)",
"max_score": 15
},
{
"name": "file:// URL in sample test",
"description": "sample_test.txt uses 'navigate to file:///workspace/orion/index.html' (file:// scheme)",
"max_score": 10
},
{
"name": "Assertion phrasing in sample test",
"description": "Both assertion steps in sample_test.txt use 'assert X is visible' without a descriptor word prefix (e.g. 'assert Welcome to Orion is visible', not 'assert the heading Welcome to Orion is visible')",
"max_score": 12
},
{
"name": "Click phrasing in sample test",
"description": "The click step uses simple 'click Learn More' or 'click the Learn More button' — but NOT a long descriptive form",
"max_score": 8
},
{
"name": "Auto-install runtime mentioned",
"description": "setup_guide.md mentions that the headless runtime is auto-installed on first use (no manual runtime install required)",
"max_score": 10
},
{
"name": "No signup/gate mentioned",
"description": "setup_guide.md does NOT describe a signup, email gate, or account creation step (Alethia has no signup requirement)",
"max_score": 7
}
]
}