From e2e-pipeline
Use when the user explicitly invokes /e2e-dispatch or asks for the E2E operations menu without specifying a particular operation. This is a unified router — for specific operations, prefer e2e-test, e2e-map, e2e-walkthrough, or e2e-skill-ops directly.
How this skill is triggered — by the user, by Claude, or both
Slash command
/e2e-pipeline:e2e-dispatchThe summary Claude sees in its skill listing — used to decide when to auto-load this skill
Route E2E operations to the correct executor with auth pre-flight.
Route E2E operations to the correct executor with auth pre-flight.
/e2e-dispatch <intent> [options]
| Argument | Example |
|---|---|
--test <flow> | /e2e-dispatch --test login-flow |
--test --tag <tag> | /e2e-dispatch --test --tag smoke |
--test --suite <name> | /e2e-dispatch --test --suite smoke |
--test --all | /e2e-dispatch --test --all |
--map <app> | /e2e-dispatch --map my-app |
--map --interactive | /e2e-dispatch --map my-app --interactive |
--walk [mapping] | /e2e-dispatch --walk admin-panel |
--flow [--from source] | /e2e-dispatch --flow --from plan.md |
--flow --smoke | /e2e-dispatch --flow --smoke |
--flow --verify-only | /e2e-dispatch --flow --verify-only login-flow |
--analyze <path> | /e2e-dispatch --analyze .claude/e2e/reports/trace.zip |
--compile [flow] | /e2e-dispatch --compile login-flow |
--compile --all | /e2e-dispatch --compile --all |
--ops [mode] | /e2e-dispatch --ops --debug |
No args or ambiguous request: present the routing menu and ask user to choose:
Available E2E operations:
--test— Run E2E test flows (single flow, suite, tag, or all)--map— Create or update UI element mappings--flow— Generate & verify E2E flows from plans/specs/PRs, or smoke test--walk— Interactive walkthrough / explore UI--analyze— Analyze a Playwright trace file--compile— Compile flow YAML to standalone bash test scripts--ops— Debug, maintain, or evaluate E2E skills--help— Interactive help guide & topic deep-dive--doc-sync— Scan docs for gaps, write updates/e2e-debug— Debug frontend runtime bugs (inject logs → browser observe → cleanup) (direct skill — not dispatched, suggest user invoke directly)Which operation? (or describe what you want to do)
Routing priority (when user intent matches multiple routes):
--flag → use that route directly--test, "record/map" → --map, "generate flow/verify flow/smoke test" → --flow, "walk/explore/browse" → --walk, "analyze/trace" → --analyze, "compile/generate script/bash script" → --compile, "fix skill/maintain skill/evaluate skill" → --ops, "help/how/what commands" → --help, "doc/sync docs/update docs" → --doc-sync, "debug runtime/inject logs/browser debug/data shape/why is X empty" → suggest /e2e-debug (direct invocation)"debug" disambiguation: The word "debug" alone is ambiguous — ask the user to clarify:
--ops --debug/e2e-debug--ops --debug/e2e-debugUnknown command (e.g., --deploy, --something): respond with "Unknown e2e operation. Available operations:" + the menu above.
Applies to: --test, --map, --flow, --walk only (browser operations).
Skipped for: --analyze, --compile, --ops, --help, --doc-sync, /e2e-debug (no browser needed).
Determine app name from the arguments:
--test: resolve flow file → read mapping: field → load mapping → app field--test --suite: resolve suite → collect all unique site mappings → app fields (check all)--map <app>: use the <app> argument directly--flow: resolve mapping from --mapping arg or auto-discover → app field--walk [mapping]: read mapping file → app fieldCheck auth profile exists:
ls ~/.agent-browser/<app>/ 2>/dev/null && echo "EXISTS" || echo "MISSING"
Check server reachable:
curl -s -o /dev/null -w "%{http_code}" <base_url>
Any 2xx or 3xx = OK. Fail → report and stop.
Profile missing or expired:
agent-browser --profile ~/.agent-browser/<app> --headed open <base_url>auth.manual_prompt (or default: "Please complete login in the browser. Let me know when done.")agent-browser get url → verify against auth.verificationagent-browser close), proceed to dispatchAuth gate pass → proceed to dispatch
Invoke Skill: "e2e-test" with the original arguments (flow name, --tag, --suite, --all, --pr, --issue, --video).
The e2e-test skill handles flow resolution and dispatches e2e-test-runner agent(s).
Invoke Skill: "e2e-map" with --scope and --page arguments.
The e2e-map skill runs codebase analysis then dispatches e2e-mapper agent.
Invoke Skill: "e2e-map" — runs entirely in main context (interactive mode).
Invoke Skill: "e2e-flow" with --from, --smoke, --verify-only, --mapping, --pr, --issue, --no-verify, --no-video arguments.
The e2e-flow skill runs codebase scan then dispatches flow-writer + flow-verifier agents.
Invoke Skill: "e2e-walkthrough" with mapping name and any --mode, --sites, --pr, --issue, --no-video arguments.
Runs entirely in main context (interactive).
Rerouting: If --walk is combined with --smoke or --verify, redirect to --flow:
--walk --smoke → --flow --smoke (smoke test is now automated via flow-writer)--walk --verify → --flow --verify-only (verification is now automated via flow-verifier)Forward all additional flags to the target route (e.g., --walk --smoke --site carlove → --flow --smoke --site carlove). The mapping argument from --walk becomes the --mapping argument for --flow.
Dispatch directly — no skill needed:
Agent(subagent_type="e2e-trace-analyzer"):
trace_path: <absolute path to trace.zip>
report_dir: <dirname of trace_path>
Present summary when agent returns.
Invoke Skill: "e2e-compile" with flow name or --all arguments.
The e2e-compile skill compiles flow YAML to standalone bash test scripts.
Invoke Skill: "e2e-skill-ops" with --debug/--maintain/--add-feature/--evaluate mode.
Invoke Skill: "e2e-help" with topic name or --feedback arguments.
Invoke Skill: "e2e-doc-sync" with --fix or --check arguments.
| Route | Default | Override |
|---|---|---|
--test | Background | --fg for foreground |
--map (no --interactive) | Background | --fg for foreground |
--map --interactive | Foreground | — |
--flow | Foreground | — |
--walk | Foreground | — |
--analyze | Background | --fg for foreground |
--compile | Foreground | — |
--ops | Foreground | — |
--help | Foreground | — |
--doc-sync | Foreground | — |
Background = run_in_background: true on Agent dispatch. Main context free for other work.
Foreground = wait for completion, interactive.
Direct invocation of any target skill (/e2e-test, /e2e-map, /e2e-flow, /e2e-walkthrough, /e2e-compile, /e2e-skill-ops, /e2e-help, /e2e-doc-sync, /e2e-debug) still works.
/e2e-dispatch is a convenience unified entry point — not required.
| Need | Command |
|---|---|
| Run one flow | /e2e-dispatch --test login-flow |
| Smoke suite | /e2e-dispatch --test --suite smoke |
| All smoke tagged | /e2e-dispatch --test --tag smoke |
| Map new app | /e2e-dispatch --map my-app |
| Update one page | /e2e-dispatch --map my-app --page dashboard |
| Generate flow from plan | /e2e-dispatch --flow --from plan.md |
| Smoke test all pages | /e2e-dispatch --flow --smoke |
| Verify existing flow | /e2e-dispatch --flow --verify-only login-flow |
| Interactive explore | /e2e-dispatch --walk admin-panel |
| Analyze trace | /e2e-dispatch --analyze .claude/e2e/reports/20260306/trace.zip |
| Compile one flow | /e2e-dispatch --compile login-flow |
| Compile all flows | /e2e-dispatch --compile --all |
| Debug skill issue | /e2e-dispatch --ops --debug |
| Record a test run | /e2e-dispatch --test login-flow --video |
| Walkthrough no video | /e2e-dispatch --walk admin-panel --no-video |
| Get help on suites | /e2e-dispatch --help suites |
| Report doc gap | /e2e-dispatch --help --feedback "suite subset" |
| Sync docs after changes | /e2e-dispatch --doc-sync --fix |
npx claudepluginhub iamcxa/kc-claude-plugins --plugin e2e-pipelineProvides a checklist for code reviews covering functionality, security, performance, maintainability, tests, and quality. Use for pull requests, audits, team standards, and developer training.