[night-shift] Implement the Dash browser front door

.codex/skills/qa-night-shift/SKILL.md

@@ -131,8 +131,9 @@ Typical flow:
    resume attempt when the run was interrupted
 10. use `night-shift resolve` or `night-shift resume` only if the run actually
    requires it
-11. use `night-shift dash` when the QA pass needs browser-visible bootstrap,
-    SSE, command, audit, or raw artifact validation
+11. use `night-shift dash` when the QA pass needs browser-visible guided init,
+    bootstrap, SSE, DAG/detail sync, command, audit, or raw artifact
+    validation
 
 For review-driven investigations, replace steps 3-4 with:
 
@@ -145,8 +146,12 @@ In review-driven runs, pay attention to repo-state evidence:
 
 - the stored open-PR snapshot captured during planning
 - whether `status`, `report`, or the dashboard show repo-state drift
+- whether `night-shift dash` guided init can move an uninitialized repo into
+  the normal initialized workspace without restarting Dash
 - whether `night-shift dash` bootstrap and SSE payloads stay aligned with the
   journal-backed repo-state snapshot
+- whether the Dash workspace shows the pending or active DAG with synchronized
+  graph, list, and detail panes wired to the persisted task metadata
 - whether `night-shift report` shows the actionable/impacted subtree and
   replacement lineage, while the persisted `report.md` remains readable
   without live GitHub refresh
@@ -180,7 +185,7 @@ In review-driven runs, pay attention to repo-state evidence:
 - whether `status`, `report`, and the dashboard agree on the confidence posture
   and its reasons
 - whether Dash command handlers mutate the same durable run state as the CLI,
-  without scraping command output
+  without scraping command output or inventing a separate control plane
 - whether `provenance` records the expected prompt paths, payload artifacts,
   verification evidence, worktree paths, and PR linkage
 - whether `doctor` classifies interrupted tasks as `safe_to_resume`,

docs/getting-started.md

@@ -176,5 +176,9 @@ If you want a dry proof that the end-to-end harness is wired correctly:
 
 ```sh
 night-shift --demo
-night-shift --demo --ui
+night-shift dash
 ```
+
+`night-shift --demo` keeps validating the fixture-backed CLI flow. Use
+`night-shift dash` when you want the real browser workspace and its live audit
+surface.

docs/providers-and-delivery.md

@@ -114,8 +114,10 @@ leave that disabled and still use the PR-body overlay.
 The local dashboard now uses `night-shift dash`. It binds to `127.0.0.1` and
 serves a Dash backend for:
 
-- structured bootstrap state for the current repository
-- SSE-first live updates
+- structured bootstrap state for the current repository, including init state,
+  selected-run audit data, DAG metadata, repo-state drift, delivery context,
+  and runtime identity
+- SSE-first live updates for the selected run and audit surface
 - browser command handlers for `init`, `plan`, `plan --from-reviews`,
   `resolve`, `start`, and `resume`
 - audit and raw artifact routes for reports, provenance, logs, payloads, and
@@ -128,7 +130,8 @@ shelling out to `night-shift` subprocesses.
 
 `night-shift --demo` exercises a fixture-backed flow and prints a compact proof
 summary. The headless demo validates `plan`, `start`, `status`, and `report`.
-The UI demo validates Dash bootstrap plus browser-command execution as well.
+The UI demo validates Dash bootstrap plus browser-command execution as well,
+but the supported human entrypoint is `night-shift dash`.
 
 Demo artifacts live under:
 

docs/run-lifecycle.md

@@ -61,13 +61,17 @@ and the next action becomes `night-shift start`.
 night-shift doctor
 night-shift resume --explain
 night-shift resume
-night-shift resume --run run-123 --ui
+night-shift dash
 ```
 
 Night Shift reloads the saved run, validates the saved environment, recovers
 in-flight tasks, and continues orchestration. It does not re-resolve provider
 or environment settings; it reuses what the run journal already saved.
 
+Use `night-shift dash` when you want the browser workflow for the same durable
+run state. Use `resume --explain` and `resume` when you prefer the CLI
+recovery path.
+
 `doctor` and `resume --explain` are the read-only recovery surfaces. They
 inspect the saved run, active lock, worktrees, logs, review drift, and
 interrupted task states, then classify each task as `safe_to_resume`,

src/night_shift/app.gleam

@@ -134,20 +134,15 @@ fn run_initialized_command(
           )
         Error(message) -> message
       })
-    types.Start(run, False) -> io.println(start(repo_root, run, config))
+    types.Start(run) -> io.println(start(repo_root, run, config))
     types.Status(run) -> io.println(status(repo_root, run, config))
     types.Report(run) -> io.println(report(repo_root, run, config))
     types.Provenance(run, task_id, format) ->
       io.println(provenance(repo_root, run, task_id, format, config))
     types.Doctor(run) -> io.println(doctor(repo_root, run, config))
     types.Resolve(run) -> io.println(resolve(repo_root, run, config))
-    types.Resume(run, False, False) ->
-      io.println(resume(repo_root, run, config))
-    types.Resume(run, False, True) -> io.println(doctor(repo_root, run, config))
-    types.Start(_, True) | types.Resume(_, True, False) ->
-      io.println(
-        "The `--ui` entrypoint was replaced by `night-shift dash`.",
-      )
+    types.Resume(run, False) -> io.println(resume(repo_root, run, config))
+    types.Resume(run, True) -> io.println(doctor(repo_root, run, config))
     _ -> io.println("Unsupported command.")
   }
 }

src/night_shift/cli.gleam

@@ -31,7 +31,7 @@ pub fn usage() -> String {
 ///
 /// ```gleam
 /// > parse(["start", "--run", "latest"])
-/// Ok(types.Start(types.LatestRun, False))
+/// Ok(types.Start(types.LatestRun))
 /// ```
 ///
 /// ```gleam
@@ -225,7 +225,7 @@ fn parse_init_flags(
 }
 
 fn parse_start(args: List(String)) -> Result(types.Command, String) {
-  parse_start_flags(args, types.LatestRun, False)
+  parse_start_flags(args, types.LatestRun)
 }
 
 fn parse_reset(args: List(String)) -> Result(types.Command, String) {
@@ -248,35 +248,31 @@ fn parse_reset_flags(
 fn parse_start_flags(
   args: List(String),
   run: types.RunSelector,
-  ui_enabled: Bool,
 ) -> Result(types.Command, String) {
   case args {
-    [] -> Ok(types.Start(run, ui_enabled))
-    ["--run", "latest", ..rest] ->
-      parse_start_flags(rest, types.LatestRun, ui_enabled)
-    ["--run", run_id, ..rest] ->
-      parse_start_flags(rest, types.RunId(run_id), ui_enabled)
+    [] -> Ok(types.Start(run))
+    ["--run", "latest", ..rest] -> parse_start_flags(rest, types.LatestRun)
+    ["--run", run_id, ..rest] -> parse_start_flags(rest, types.RunId(run_id))
     [flag, ..] -> Error("Unsupported start flag: " <> flag)
   }
 }
 
 fn parse_resume(args: List(String)) -> Result(types.Command, String) {
-  parse_resume_flags(args, types.LatestRun, False, False)
+  parse_resume_flags(args, types.LatestRun, False)
 }
 
 fn parse_resume_flags(
   args: List(String),
   run: types.RunSelector,
-  ui_enabled: Bool,
   explain_only: Bool,
 ) -> Result(types.Command, String) {
   case args {
-    [] -> Ok(types.Resume(run, ui_enabled, explain_only))
+    [] -> Ok(types.Resume(run, explain_only))
     ["--run", "latest", ..rest] ->
-      parse_resume_flags(rest, types.LatestRun, ui_enabled, explain_only)
+      parse_resume_flags(rest, types.LatestRun, explain_only)
     ["--run", run_id, ..rest] ->
-      parse_resume_flags(rest, types.RunId(run_id), ui_enabled, explain_only)
-    ["--explain", ..rest] -> parse_resume_flags(rest, run, ui_enabled, True)
+      parse_resume_flags(rest, types.RunId(run_id), explain_only)
+    ["--explain", ..rest] -> parse_resume_flags(rest, run, True)
     [flag, ..] -> Error("Unsupported flag: " <> flag)
   }
 }