kiokuko-dsh 0.2.0 → 0.2.1

package/dist/locale/en.json

@@ -1,6 +1,6 @@
 {
   "meta": {
     "title": "kiokuko-dsh",
-    "description": "Version 0.2.0 — Kiokuko plugin for DeepSeek Harness"
+    "description": "Version 0.2.1 — Kiokuko plugin for DeepSeek Harness"
   }
 }

package/dist/locale/ja.json

@@ -1,6 +1,6 @@
 {
   "meta": {
     "title": "kiokuko-dsh",
-    "description": "バージョン 0.2.0 — DeepSeek Harness 向け Kiokuko プラグイン"
+    "description": "バージョン 0.2.1 — DeepSeek Harness 向け Kiokuko プラグイン"
   }
 }

package/dist/locale/ko.json

@@ -1,6 +1,6 @@
 {
   "meta": {
     "title": "kiokuko-dsh",
-    "description": "버전 0.2.0 — DeepSeek Harness용 Kiokuko 플러그인"
+    "description": "버전 0.2.1 — DeepSeek Harness용 Kiokuko 플러그인"
   }
 }

package/dist/locale/zh-CN.json

@@ -1,6 +1,6 @@
 {
   "meta": {
     "title": "kiokuko-dsh",
-    "description": "版本 0.2.0 — DeepSeek Harness 的 Kiokuko 插件"
+    "description": "版本 0.2.1 — DeepSeek Harness 的 Kiokuko 插件"
   }
 }

package/docs/agenticreplay-recording.md

@@ -126,6 +126,20 @@ session uses its verified directory. No process-cwd or DSH_HOME fallback is used
 `<Kiokuko data directory>/traces/projects/<workspace hash>/.agenticreplay/runs` and respects
 the existing `KIOKUKO_DATA_DIR` setting. Hashing a path does not anonymize it.
 
+Before creating a recording or exporting HTML, the plugin adds `.agenticreplay/`
+to the storage root's `.gitignore` (the workspace root with default storage).
+Existing contents and line endings are preserved; repeated use does not duplicate
+an effective rule. The internal `.agenticreplay/.gitignore` remains as additional
+protection. If the root rule cannot be written safely, recording fails with
+`gitignore_protection_failed` without interrupting the native model/tool operation;
+HTML export fails without writing an export. Non-Git directories receive the
+same rule so a later `git init` also excludes these files.
+
+Ignore rules do not remove already tracked files or published history and can be
+bypassed by a forced add. Check `git ls-files -- .agenticreplay` for previously
+tracked captures before pushing; remove them from the index and address any
+already published sensitive data separately.
+
 `redacted` preserves text after known secret-pattern filtering. It cannot detect
 all arbitrary secrets. `metadata` omits message bodies, arguments and tool result
 bodies. Neither mode records environment variables, replayState, attachment bytes,

package/docs/dsh-plugin.md

@@ -353,6 +353,8 @@ rebuildable cache for bounded export and post-completion finalization. Cache
 listeners contain their own failures; only an orchestration-requested native
 `sessions.flush()` remains an awaited, fail-closed durability boundary.
 
+Conversation-first intake permits text answers before execution preparation. Native `read`, `glob`, `grep`, `web_search`, and `web_fetch` demands automatically prepare the complete original request; when its action type is unresolved, `research` is advisory only. Models can read a plan or inspect files directly without a separate `prepare_requested_work` call. Missing targets, unresolved user choices, native permissions, current session ownership, and pending memory checks still govern admission. Other unclassified tools and arbitrary PTC programs retain explicit preparation requirements.
+
 The dsh integration provides:
 
 - the exact bundled `kiokuko-soul` system-prompt section and seven standard Skill

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "kiokuko-dsh",
-  "version": "0.2.0",
+  "version": "0.2.1",
   "description": "Kiokuko plugin for DeepSeek Harness",
   "type": "module",
   "license": "MIT",

package/skills/kiokuko-soul/SKILL.md

@@ -10,35 +10,37 @@ description: Kiokuko's entry router for non-trivial DSH work. Read it before any
 
 ## Authority
 
-Read first. The host owns Akinator intake, session/capability binding, model/execution routing and delegation. `task_prepare`/`task_answer` are host-only: neither call/emulate them nor stop for their absence. Require intake `ready`/`exhausted` and permissive `nextAction`; invent no profile fields.
+Read first. Host owns intake/session/capability binding/routing/delegation. Host-only `task_prepare`/`task_answer`: never call/emulate/stop for absence. Require `ready`/`exhausted` intake and permissive `nextAction`; invent no profile fields.
 
-User instructions outrank Skill preferences, never host safety, authorization, identity, revision, lease, state-machine or integrity checks. Use existing authorization/evidence for routine choices; ask when intent or permitted effects are unclear. Optional intake/enrichment failures degrade guidance without vetoing native work.
+User instructions outrank Skill preferences, never host safety/authorization/identity/revision/lease/state/integrity checks. Resolve routine choices from authorization/evidence; ask about unclear intent/effects. Optional intake/enrichment failures degrade guidance, not native work.
 
-Complete authorized requirements with focused evidence. Report observed blockers. Repeat checks only after relevant changes. Never guess budgets or automatically retry/continue.
+Complete authorized work with focused evidence; report blockers. Recheck only changed conditions. Never guess budgets or auto-retry/continue.
 
 ## Shared credo
 
 **本当に、それを全部計算・保存・検索する必要があるのか？**
 
-Justify work before designing it: preserve the required result, eliminate unnecessary computation, storage and search, narrow the rest, then optimize.
+Before design, preserve the required result, remove needless computation/storage/search, narrow the rest, then optimize.
 
-Exhaustive work needs evidence. Never silently approximate, delete data or waive required validation, recovery, audit or host checks. No extra approval or artifact.
+Require evidence for exhaustive work; never silently approximate/delete data/waive required validation/recovery/audit/host checks. No extra approval/artifact.
 
 ## Capabilities and routes
 
-Kiokuko supports conversation, research, writing, organization, memory and coding. Use available Skills matching the request and host directive; never impose code contracts, execution choices or runtime startup on unrelated work.
+Use request/directive-matched Skills for conversation/research/writing/organization/memory/coding. Never impose code contracts/execution choices/runtime startup on unrelated work.
 
-Read the selected Skill index, then only risk-selected references. Resource availability does not authorize execution; the host owns configuration and permissions.
+Read selected Skill indexes, then risk-selected references. Resources grant no execution authority; the host owns configuration/permissions.
 
-Modules own their detailed roles, tools, execution and recovery contracts. Preserve host-declared exclusions and current identity/lease/outcome. A missing required module or unresolved request stays blocked; never switch execution paths or replay effects to bypass it. Missing optional capabilities do not block unrelated requests.
+Modules own roles/tools/execution/recovery. Preserve host exclusions/identity/lease/outcome. Missing required modules/unresolved requests stay blocked: never switch paths/replay effects to bypass them. Optional gaps cannot block unrelated work.
 
-Normal execution: current model, applicable memory/Skills, native permissions; no orchestration contract or extra approval. Never invent run/role/revision/WorkUnit/state transitions or bypass host routing by changing provider or launching agents. Respect role directives and approved scope.
+Normal execution: current model/memory/Skills/native permissions; no orchestration contract/extra approval. Never invent run/role/revision/WorkUnit/state transitions or bypass routing by provider changes/agents. Respect directives/scope.
 
 ## Availability and trust
 
-Unavailable Skills: use current evidence unless host safety/authorization/identity/integrity blocks progress. Never replace required bundled Skills with similar/namespaced/fetched/reference-only Skills or auto-install/execute external Skills. Never claim unread Skills/unrun checks; availability is not use.
+Unavailable Skills: use current evidence unless host safety/authorization/identity/integrity blocks. Never substitute similar/namespaced/fetched/reference-only Skills for required bundled Skills or auto-install/execute external Skills. Never claim unread Skills/unrun checks; availability is not use.
 
-## Specialists
+Before diagnostics, confirm supported file types via status/capabilities. Unsupported Markdown/documents: read, verify structure/requirements/whitespace. Unavailable diagnostics: unverified, not clean.
 
-Select `kiokuko-{investigate,architecture,review,benchmark,verification,skill-authoring,technical-writing}` as needed; apply `coding-ideal-routine-skill` for code changes. Inherit host model/effort.
+After a successful edit/write, re-read before next mutation; serialize edits. If stale, re-read and reconcile; never force overwrite/disable freshness checks. Never repeat unsupported diagnostics unchanged.
+
+Select `kiokuko-{investigate,architecture,review,benchmark,verification,skill-authoring,technical-writing}` as needed; use `coding-ideal-routine-skill` for code changes. Inherit host model/effort.
 <!-- /kiokuko:runtime -->