diff --git a/docs/active-work.md b/docs/active-work.md
index 147f568..625509f 100644
--- a/docs/active-work.md
+++ b/docs/active-work.md
@@ -414,13 +414,25 @@ compatible.
 - **DAP waits for Stage 2, not Stage 1** — that dependency is now
   satisfied.
 
-## Test ambient-root isolation — FRAMING OPEN, revision 4
+## Test ambient-root isolation — IMPLEMENTATION OPEN
 
-- **Branch `test-ambient-config-isolation`**, worktree
-  `../pmacs-test-isolation`, based on `githubsucks/main` @ `4cd4a7b`.
-  **Framing only; no code, no PR yet.**
-  `docs/test-ambient-config-isolation-framing.md` revision 4, three review
-  rounds closed (eight blocking, six major, all accepted).
+- **Framing MERGED (#201)**;
+  `docs/test-ambient-config-isolation-framing.md` revision 4, three
+  review rounds closed (eight blocking, six major, all accepted).
+- **Implementation branch `test-ambient-isolation-impl`**, worktree
+  `../pmacs-test-isolation-impl`, based on `githubsucks/main` @
+  `54a092e`. The framing-only worktree `../pmacs-test-isolation`
+  (branch `test-ambient-config-isolation`) is spent — its doc is on
+  `main`.
+- **What landed on the branch.** `pmacs::bootstrap::BootstrapRoots`
+  (config/data/state/cache, `ambient()` for production), reachable from
+  `EditorState::new_with_roots` **and** `open_with_roots` and consulted
+  again by `install_state_dirs`; all 342 in-process construction sites
+  in 65 files migrated; `journey_acceptance` keeps the ambient `open`
+  and re-execs itself per test with controlled roots instead; the shared
+  daemon and PTY spawners now set all five storage variables;
+  `tests/ambient_isolation_acceptance.rs` carries the hostile-environment
+  proof and the adoption ratchet.
 - **What it is.** Integration tests use the developer's real ambient
   roots. `#[cfg(not(test))]` guards config loading against the crate's
   own unit tests only, so the **65** files in `tests/` that construct an
@@ -455,9 +467,9 @@ compatible.
 
   ```sh
   git fetch githubsucks
-  git worktree add ../pmacs-test-isolation \
-    -b test-ambient-config-isolation \
-    githubsucks/test-ambient-config-isolation
+  git worktree add ../pmacs-test-isolation-impl \
+    -b test-ambient-isolation-impl \
+    githubsucks/test-ambient-isolation-impl
   ```
 
 ## Reap-ledger silent failures — MERGED (#202); kept for its parked follow-ons
diff --git a/docs/agent-handoff.md b/docs/agent-handoff.md
index d4542f4..ad61320 100644
--- a/docs/agent-handoff.md
+++ b/docs/agent-handoff.md
@@ -1639,6 +1639,35 @@ git diff --check
 Machine-specific caveats — re-verify on a machine you haven't used
 before trusting them:
 
+- **Ambient storage roots: control all FIVE, not four.** Until the
+  ambient-root isolation lane lands
+  (`docs/test-ambient-config-isolation-framing.md`), the ~96 integration
+  suites read the developer's real `~/.config/pmacs/init.lua` and
+  **write** bundled packages into the real data root:
+  `#[cfg(not(test))]` guards the crate's own unit tests only, and
+  `EditorState::new` materializes packages outside every `cfg` guard.
+  A local full-suite run therefore needs, all pointed at a fresh
+  directory:
+
+  ```
+  XDG_CONFIG_HOME  XDG_DATA_HOME  XDG_STATE_HOME  XDG_CACHE_HOME
+  PMACS_STATE_HOME
+  ```
+
+  **The fifth is not redundant.** `PMACS_STATE_HOME` outranks
+  `XDG_STATE_HOME` (`src/state.rs`), so redirecting the four XDG
+  variables on a machine that exports it leaves the real state root
+  live. `HOME` is deliberately left alone: it is the fallback for the
+  XDG roots (already covered once they are set) and separately drives
+  `~`-expansion, which `find_file_acceptance` pins on purpose.
+  A run isolating only `XDG_CONFIG_HOME` stops the `init.lua` reads and
+  still writes through the real data root — every local gate run in this
+  repo before 2026-07-31 had that hole.
+
+  **After the lane lands** the in-process population isolates itself
+  through `EditorState::new_with_roots(&crate::iso::roots())` and the
+  five variables are belt-and-braces rather than required.
+
 - **basedpyright**: the desktop binary was **never broken** — this was a
   real code defect, diagnosed and fixed 2026-07-29 (see §5, "A `Drop`
   body runs before its fields"). `RuntimeHandles::drop` joined its reader