feat(onboard): recommendations store with once-semantics and self-naming bootstrap (#110173)

* feat(onboarding): persist app recommendations once

* feat(onboarding): add self-naming birth sequence

* chore(wizard): hoist bootstrap-defer imports to the top of the module

* docs(bootstrap): persist agreed identity into IDENTITY.md and SOUL.md as well

* docs: regenerate docs map after onboarding plan merge

* fix(wizard): reuse a pending stored offer instead of rescanning apps

* fix(onboard): deduplicate bootstrap recommendations

docs/start/bootstrapping.md

@@ -18,8 +18,16 @@ On the first run against a brand-new workspace (default `~/.openclaw/workspace`)
 OpenClaw:
 
 - Seeds `AGENTS.md`, `SOUL.md`, `TOOLS.md`, `IDENTITY.md`, `USER.md`, `HEARTBEAT.md`, and `BOOTSTRAP.md`.
-- Has the agent follow `BOOTSTRAP.md`: a free-form conversation (not a fixed Q&A form) to settle on a name, personality, and vibe.
-- Writes what it learns into `IDENTITY.md`, `USER.md`, and `SOUL.md`.
+- Has the agent follow a capped three-beat birth sequence: it proposes its own
+  name, shares one short soul/vibe line, and asks whether you want the minimal
+  recommended plugin set or maximum convenience.
+- Persists the agreed identity twice: into `IDENTITY.md` and `SOUL.md` (what the
+  agent reads about itself) and via `openclaw agents set-identity` (what channels
+  and the UI display).
+- Reads app recommendations already stored during onboarding without rescanning.
+  Official plugins use `openclaw plugins install <id>`; third-party ClawHub
+  skills remain explicit opt-ins. After the choice is handled, the agent
+  acknowledges the stored offer so it never asks again.
 - Deletes `BOOTSTRAP.md` once the workspace looks configured, so the ritual only runs once.
 
 A workspace counts as configured once `SOUL.md`, `IDENTITY.md`, or `USER.md` has

docs/start/onboarding-redesign.md

@@ -186,16 +186,21 @@ Remote-gateway onboarding keeps its legacy conversational handoff
 ### Phase 5 — hatch and bootstrap (planned)
 
 - Custodian creates a nameless agent (tool call); the agent's bootstrap opens
-  with self-naming and a self-drawn avatar (image-gen ladder: model-generated
-  candidates → preset marks → keep logo). Same thread, avatar swap; the claw
-  mark stays reserved for the custodian. Cap the birth sequence at roughly
-  three beats (name+face → soul line → skills question) before the agent is
-  useful.
+  with self-naming. PR1 ships the ceremony capped at three beats (name → soul
+  line → skills question) and defers the self-drawn avatar/image-gen ladder
+  (model-generated candidates → preset marks → keep logo) to a follow-up. Same
+  thread, avatar swap; the claw mark stays reserved for the custodian. The
+  agreed identity persists twice: into `IDENTITY.md`/`SOUL.md` (what the agent
+  reads) and via `openclaw agents set-identity` (what channels and the UI
+  display).
 - Recommendations (phase 1 service, stored scan with once-semantics) land as
   the last bootstrap step before the bootstrap file is removed: "minimal set
-  or maximum convenience?" Channel connect buttons carry per-channel setup
-  playbooks; the agent collects credentials conversationally and relays config
-  writes to the custodian ("asking OpenClaw…" is the canonical idiom).
+  or maximum convenience?" The bootstrap reads the stored offer via
+  `openclaw onboard recommendations --json` (opaque install IDs only) and
+  acknowledges it after the choice is handled so it never asks again. Channel
+  connect buttons carry per-channel setup playbooks; the agent collects
+  credentials conversationally and relays config writes to the custodian
+  ("asking OpenClaw…" is the canonical idiom).
 - Self-learning is asked, not announced, and doubles as skill-workshop
   consent; describe ClawHub's release-trust, scan, verification, and integrity
   checks plus the publisher-code warning — never imply every release is signed.