| name | e2e |
| description | Domain-agnostic AI-native scenario runner (e2e-agent CLI). Scenarios are markdown prompts; evidence is written by the CLI. Use when the user invokes /e2e or asks to run a live scenario test and collect evidence under .e2e/results/. Agents use only e2e-agent commands — never import internal framework modules. |
/e2e — e2e-agent (domain-agnostic)
This skill is fully generic. It does not know SAP, MCP, or any particular service. Your project supplies a config + adapter; see SPEC.md.
Agents use only the CLI below. Do not import scripts/framework/*. Do not generate per-scenario TypeScript.
Commands
e2e-agent list [--config <path>] [suite]
e2e-agent show <code> --config <path>
e2e-agent run <code> --config <path> [--param value ...]
e2e-agent dispatch <code> --config <path> --acp --agent <acp-id> [--param value ...]
e2e-agent help
Config defaults via E2E_CONFIG or e2e.config.yaml at repo root when omitted.
OpenADT (this repo)
bun run e2e -- list
bun run e2e -- show ls-1
bun run e2e -- run ls-1 --destination ABC
bun run e2e -- run mcp-1 --destination ABC
Project profile: e2e.config.yaml. Product spec: specs/mcp-ai-testing.md.
On /e2e <code> [params]
- Map user params to CLI flags (e.g.
--destination ABC or full ABC_200_USER_EN).
- Run
e2e-agent run <code> --config <project-config> … (or dispatch with --acp).
- Read stdout: exit code,
E2E_EVIDENCE_FILE=….
- Report PASS/FAIL, Given/When/Then summary, assertion highlights from evidence.
| User intent | Command |
|---|
| Run locally (default) | e2e-agent run … |
| Delegate (no local service spawn) | e2e-agent dispatch … --acp --agent <id> |
Report to user
Local run: PASS/FAIL, evidence path, Given/When/Then, per-step assertion table (quote failures with expected vs actual).
ACP dispatch: E2E_DISPATCH_FILE, target agent, pasteable prompt and command.local. Do not claim PASS/FAIL until external agent returns evidence.
Configuration
Project config lives in e2e.config.yaml at the repo root (adapter, suites, optional autoclean). See SPEC.md for framework flags, scenario format, adapter contract, and evidence shape.
