@zq-silk/yui 0.7.1 → 0.8.2

package/skills/yui-leader/SKILL.md

@@ -296,9 +296,12 @@ Choose before creating the WorkItem:
   credentials, user-owned independent Session, durable lifecycle, or repeated
   dispatches to a Task-bound Worker instance.
 
-Keep review execution separate from implementation. A reviewer uses the single
-built-in write-capable `reviewer` Profile, but Yui grants that capability only
-inside a fresh ReviewRound-owned worktree created from its exact frozen scope:
+Keep review execution separate from implementation. No global Reviewer is
+required: when review is disabled, inspect and decide directly or delegate a
+bounded review to a native subagent or ordinary Worker. When a managed
+ReviewRound is explicitly configured, its reviewer uses the single built-in
+write-capable `reviewer` Profile, but Yui grants that capability only inside a
+fresh ReviewRound-owned worktree created from its exact frozen scope:
 the assigned WorkItem Candidate or the committed Integration heads of a
 Task-final Review. Never reuse the Candidate/Worker workspace or its
 implementation Role Session. Codex and Claude may use their normal configured
@@ -319,7 +322,7 @@ Before the first delegated WorkItem, or after the Profile catalog changes,
 inspect the available Profiles:
 
 ```sh
-yui profile list
+yui config profile list
 ```
 
 Choose the Profile by the work's meaning. `worker`, `implementer`, and
@@ -375,7 +378,7 @@ use `worker`. A Profile is required for this path:
 
 ```sh
 yui task work update <work-id> running
-yui profile show <worker|explorer|implementer|reviewer|profile-id>
+yui config profile show <worker|explorer|implementer|reviewer|profile-id>
 ```
 
 Read the selected Profile and incorporate all applicable portable constraints
@@ -495,6 +498,10 @@ After any Candidate is submitted, inspect its exact policy, Run result,
 ReviewRounds, checks, and workspace through the exact Run Context Pack and its
 authorized expansions.
 
+- No configured review policy: review the Candidate directly, or delegate a
+  bounded evidence-gathering review to a native subagent or ordinary Worker,
+  then make the Leader-owned accept/reject decision. Do not create a Reviewer
+  Role merely to satisfy an old setup convention.
 - `always`: wait for the automatically requested ReviewRound to become
   terminal. Never bypass an active round.
 - `leader`: decide whether the existing evidence is sufficient. Request Agent

package/skills/yui-operator/SKILL.md

@@ -1,6 +1,6 @@
 ---
 name: yui-operator
-description: Route multi-project user requests into Yui Tasks, preserve durable intent, present progress, answer inputs, and administer lifecycle without taking over Leader decisions.
+description: Route multi-project user requests into Yui Tasks, configure Yui through confirmed conversation, preserve durable intent, present progress, answer inputs, and administer lifecycle without taking over Leader decisions.
 ---
 
 # Yui Operator
@@ -46,6 +46,44 @@ mind and avoid imposing a fixed heading, field, section, or character
 template; one semantic event should have one concise summary unless a later
 role adds a genuinely new decision or impact.
 
+## Configure Yui through conversation
+
+Treat configuration as an Operator-owned conversation, not a list of commands
+the user must discover or run. Start every configuration discussion by reading
+both the complete effective state and Yui's configuration catalog:
+
+```sh
+yui --json config show
+yui --json config describe
+yui --json config describe <system|runtime|workflow|resources|tools|agent|role|profile|completion>
+```
+
+Consume the top-level `data` field. Explain the relevant current values, what
+each setting changes, its accepted values or referenced records, and when the
+change takes effect. Distinguish stored values from effective defaults and say
+when a live Role Session must be stopped and relaunched. Do not infer choices
+from old setup conventions: Review, Worker, Profiles, and completion may
+intentionally be absent after the minimum setup. Leader is part of the minimum
+Task-execution configuration and must be present.
+
+For Agent-dependent Role choices such as model, effort, provider permission,
+search, settings source, or service tier, also read
+`yui --json config agent capabilities <agent-id>`. Treat that live-or-cached
+catalog and its freshness warnings as the choice authority; never invent a
+provider value from memory.
+
+When the user wants a change, narrow the discussion to the affected domains,
+present the exact before/after behavior and material consequences, and obtain
+confirmation before mutating configuration. Then perform only the confirmed
+`yui config ...` commands yourself, read `yui --json config show` back, and
+when the catalog says a Controller restart is required, include that impact in
+the confirmation and run `yui controller restart` after saving. Report the
+verified result. Never make the user execute mechanical CLI steps,
+never parse human tables, never expose secret environment values, and never
+silently create a Reviewer or enable global review. A Leader may review work
+directly or delegate review to an ordinary Worker unless the user explicitly
+configures a review Role and policy.
+
 ## Route across Projects and Tasks
 
 Inspect the catalog, Tasks, and global input Inbox before routing:
@@ -192,11 +230,11 @@ When the user requires a specific Leader or Worker provider, inspect Roles
 before routing:
 
 ```sh
-yui profile list
-yui profile show <profile>
-yui role list
-yui role show leader
-yui role show worker
+yui config profile list
+yui config profile show <profile>
+yui config role list
+yui config role show leader
+yui config role show worker
 yui task role list <task-id>
 yui task work list <task-id>
 yui task integration list <task-id>

package/skills/yui-runtime/SKILL.md

@@ -41,7 +41,7 @@ For a global Operator or custom GlobalRole Session, load the stable exact view
 before routing or acting:
 
 ```sh
-"$YUI_SESSION_CLI" role context "$YUI_ROLE" --json
+"$YUI_SESSION_CLI" session context "$YUI_ROLE" --json
 ```
 
 Global context grants no Task implementation workspace. Read a Task only after