dsh-gitbash-shell 0.5.4 → 0.6.0

package/README.md

@@ -1,5 +1,7 @@
 # dsh-gitbash-shell
 
+[![Awesome DSH Plugin](https://awesome-dsh-plugin.com/badge.svg)](https://awesome-dsh-plugin.com)
+
 > 让 DeepSeek Harness (dsh) 在 Windows 上**全部使用 Git Bash** 的官方风格插件
 > —— 以 Git for Windows 的 `bash.exe` 替换 PowerShell 执行器,并为所有模式
 > 物化对应的 Git Bash 版 agent preset。
@@ -24,6 +26,17 @@
 Git Bash host 下会拿到"暗示 PowerShell 语法的工具",请改用上面的变体;
 已装 dsh-ptc-cordis-preset 的话,`PTC 创造模式` 用户 preset 不受影响。
 
+### dsh 版本双适配(0.1.1 与 0.1.2+)
+
+dsh 0.1.2 把内置 `code` preset 改名为 `ptc`(`mode: code` → `mode: ptc`,官方不做
+兼容别名),并给各内置 preset 新增 `command-goal` 等行。本插件为受影响的变体
+(standard/code/cordis)**同时携带两个 era 的已提交组合文本**,启动时探测内置
+roster 自动选择,并记进 `.plugin-managed.json` 的 `base` 字段;探测翻转(dsh
+升级前后)自动重物化。`minimal-gitbash` 的内置底稿跨版本未变,单文本服务两个
+era。**preset id 保持 `code-gitbash` 不变**(会话钉在 id 上,改名会让已固定的
+会话报 preset not found)。无论先升级插件还是先升级 dsh,都会自动收敛;用户改
+过的目录照旧不碰。
+
 ## 安装(公开 npm 插件,推荐)
 
 npm: [dsh-gitbash-shell](https://www.npmjs.com/package/dsh-gitbash-shell)，

package/README_EN.md

@@ -1,5 +1,7 @@
 # dsh-gitbash-shell
 
+[![Awesome DSH Plugin](https://awesome-dsh-plugin.com/badge.svg)](https://awesome-dsh-plugin.com)
+
 > Run **every agent shell command through Git Bash** on Windows with DeepSeek
 > Harness (dsh) — replaces the PowerShell executor and materializes Git Bash
 > variants of all four agent presets.
@@ -14,6 +16,19 @@ dsh plugin --profile web add dsh-gitbash-shell
 `dsh.bundle` declaration, appends the package to `dsh.profile.bundles`.
 **Restart the profile's host to activate.**
 
+## Works with both dsh 0.1.1 and 0.1.2+
+
+dsh 0.1.2 renamed the built-in `code` preset to `ptc` (`mode: code` → `mode: ptc`,
+no compatibility aliases) and added new built-in rows (`command-goal`, …). For
+the affected variants (standard/code/cordis) this plugin **ships both committed
+era texts**, probes the built-in roster at every boot, records the choice in
+`.plugin-managed.json` (`base`), and re-materializes automatically when the
+detection flips. `minimal-gitbash`'s built-in base did not change across the
+rename, so one text serves both eras. **Preset ids never change**
+(`code-gitbash` keeps its historical id — sessions are pinned to ids, and a
+rename would orphan them). Either upgrade order converges automatically; directories
+you modified are still never touched.
+
 ## What it does
 
 The bundle patch (`cordis.patch.yml`):

package/assets/code-gitbash/agent.cordis.ptc.yml

@@ -0,0 +1,273 @@
+# ERA: dsh >= 0.1.2 text (built-in preset id `ptc`, presented with mode: ptc).
+# The dsh <= 0.1.1 era text (built-in id `code`) lives in agent.cordis.yml; the
+# plugin probes the roster at materialization and writes the picked text as the
+# preset's own agent.cordis.yml.
+# The `ptc` agent preset: the standard coding agent, presented as PTC mode.
+#
+# Everything in `standard` is here unchanged. What is added is the `tool-presentation`
+# row: instead of one tool call per action, the model writes a TypeScript
+# program against a generated SDK and `run_code` executes it, so a sequence
+# that would be five round trips becomes one.
+#
+# The registry itself stays on the host plane — the agent loop's scheduler and
+# the API proxy's presenters are its consumers — so what this preset owns is
+# the PRESENTATION of that registry for this agent alone. Native sessions run
+# beside this one in the same process, each seeing its own catalog.
+#
+# This file is an AGENT-PLANE composition. It is mounted under one agent's
+# scope context, so every tool and prompt section it registers belongs to that
+# session alone. The host composition (`base.cordis.yml` + `web.cordis.yml`)
+# keeps everything a preset must not own: the registries themselves, the
+# sandbox and approval stack, persistence, and the model route.
+#
+# A service row here MUST sit inside a group carrying an `isolate` realm.
+# Without one it publishes into the root realm, where it is process-global
+# rather than per-session and the second session mounting this preset collides
+# with the first; `dsh-agent-presets` rejects that at mount. `true` means an
+# entry-local realm — one private instance per mounted session, which is the
+# default this deployment wants. A shared label would instead pool one instance
+# across every session naming it.
+
+# ── identity ────────────────────────────────────────────────────────────────
+
+# The preset's own persona, shadowing the deployment default for this agent.
+# `{{model}}` and `{{cwd}}` resolve from the agent's own route and workspace.
+- id: persona
+  name: '@deepseek-ai/dsh-persona'
+  config:
+    text: >-
+      You are a coding agent powered by the {{model}} model. Your working directory is {{cwd}}.
+
+- id: agent-instructions
+  name: '@deepseek-ai/dsh-agent-instructions'
+  config:
+    maxBytes: 65536
+
+# ── shell ───────────────────────────────────────────────────────────────────
+
+# `shell-env` stays in the HOST composition: `apps/cli/src/web.ts` injects it to
+# publish `DSH_WEB_URL`/`DSH_WEB_MODE`, and a host row that injects a service is
+# the criterion for host-plane ownership — injection resolves before any session
+# exists, so there is no agent to key by. Behind a preset realm those variables
+# never reached the model's shell at all. Both shell tools consume the host
+# registry from here; the host executor is the Git Bash provider (`dsh-gitbash-shell`)
+# host-plane too.
+- id: tool-bash
+  name: '@deepseek-ai/dsh-tool-bash'
+  # Git Bash on Windows: the host executor is dsh-gitbash-shell (see $DSH_HOME/cordis.patch.yml).
+  disabled: false
+
+- id: tool-pwsh
+  name: '@deepseek-ai/dsh-tool-pwsh'
+  disabled: true
+
+# ── filesystem ──────────────────────────────────────────────────────────────
+
+# Both register into the host `tools` registry and provide nothing, so
+# they need no realm. The `fs` service and its policy stay in the host.
+- id: tool-fs
+  name: '@deepseek-ai/dsh-tool-fs'
+
+- id: tool-fs-search
+  name: '@deepseek-ai/dsh-tool-fs-search'
+  config:
+    sampleOverCapGlobResults: false
+
+# ── background jobs ────────────────────────────────────────────────────────
+
+# Only the model-facing controls. The task REGISTRY stays on the host plane:
+# its producers sit outside any realm this file could put it in — `tool-bash`
+# above resolves it with `ctx.get`, and an entry-local realm here is invisible
+# to every sibling row, so `run_in_background` would answer "background jobs
+# unavailable" while these controls sat in the catalog. The registry is keyed by
+# owning agent anyway, so one host instance serves every session. What a preset
+# chooses is whether its agent can collect and stop background work at all.
+- id: tool-jobs
+  name: '@deepseek-ai/dsh-tool-jobs'
+
+# ── skills ──────────────────────────────────────────────────────────────────
+
+# The skill REGISTRY lives in the host composition and is layered per scope:
+# these rows register into THIS preset's layer of it, so they need no realm.
+# `skill-filesystem` contributes local-root discovery for agents on this preset, and
+# `tool-skill` gives them the catalog and loader; the merged catalog also
+# carries whatever the deployment registered globally (repository plugins).
+- id: skill-filesystem
+  name: '@deepseek-ai/dsh-skill-filesystem'
+
+- id: tool-skill
+  name: '@deepseek-ai/dsh-tool-skill'
+
+# ── goals ───────────────────────────────────────────────────────────────────
+
+# The goal service and session driver stay on the host plane, where the Gateway
+# can resolve them. The human command and model-facing tool register into this
+# preset's scoped layers.
+- id: command-goal
+  name: '@deepseek-ai/dsh-command-goal'
+
+- id: tool-goal
+  name: '@deepseek-ai/dsh-tool-goal'
+
+# ── plan mode ───────────────────────────────────────────────────────────────
+
+# Plan state is per-agent by nature, so an entry-local realm is not a
+# workaround here — it is the correct lifetime.
+- id: planning
+  name: cordis:group
+  group: true
+  isolate:
+    planMode: true
+  config:
+    - id: plan-mode
+      name: '@deepseek-ai/dsh-plan-mode'
+      config:
+        section: |
+              You are in plan mode. Stay in plan mode until exit_plan_mode succeeds or the user switches the session mode. Imperative language to implement changes means plan the implementation, not execute it. A user's conversational agreement — including an answer confirming something you asked — approves nothing and does not end plan mode; fold the confirmed decision into the plan and submit it through exit_plan_mode.
+
+              Explore first. Use non-mutating reads, searches, static analysis, and checks to ground the plan in the actual repository. Do not edit or write files, change configuration, run formatters or code generation that rewrites tracked files, commit, or otherwise carry out the plan. Prefer existing functions and patterns over new machinery.
+
+              The tool catalog stays the same across modes for request-cache stability. These plan-mode rules override any later tool description or guidance that suggests using mutation tools; those tools remain listed to keep the tool catalog unchanged. Do not use todo_write to track this planning phase: it tracks implementation after an approved plan, while the plan itself belongs in exit_plan_mode.
+
+              Resolve discoverable facts by inspection. Use ask_user_question only for user-owned choices or material ambiguity that inspection cannot answer. Do not ask the user where code lives or how current behavior works when you can find out.
+
+              Make the plan decision-complete: state the goal and success criteria; group implementation changes by subsystem; identify public API, schema, and data-flow changes; cover edge cases, failure modes, tests, acceptance criteria, and explicit assumptions. Keep it concise enough to review but detailed enough that another engineer can implement it without making design decisions.
+
+              When ready, call exit_plan_mode with the complete plan markdown, starting with a # title. Make exit_plan_mode the only and final tool call in that assistant response: it presents the plan for approval, and implementation begins only in a later step after approval. Do not paste the final plan as a plain reply or ask "should I proceed?" through prose or ask_user_question. If review rejects it, incorporate the feedback and present again. If the review channel is unavailable or aborted, stay in plan mode and ask the user to switch modes manually; do not proceed with implementation.
+
+# ── compaction ──────────────────────────────────────────────────────────────
+
+# `compaction-basic` reads `toolResultPrune` through `ctx.get`, so the pruner must
+# share this realm rather than sit outside it.
+#
+# `tokenMeter` is deliberately NOT in this realm: the meter stays on the HOST
+# plane, and the rows here resolve that one instance. It takes no configuration,
+# keys every fold by Session, and owns the context-meter projection units the
+# browser reads for every session — behind a realm those units would come and go
+# with whichever presets happen to be mounted. What a preset chooses is whether
+# its agent compacts at all, which is `compaction-basic` below.
+- id: compaction
+  name: cordis:group
+  group: true
+  isolate:
+    compaction: true
+    toolResultPruner: true
+  config:
+    - id: compaction-basic
+      name: '@deepseek-ai/dsh-compaction-basic'
+
+    - id: command-compact
+      name: '@deepseek-ai/dsh-command-compact'
+
+    - id: tool-result-pruner
+      name: '@deepseek-ai/dsh-compaction-tool-result-pruner'
+      config:
+        thresholdChars: 8192
+        headChars: 4096
+        tailChars: 1024
+
+# ── delegation and workflows ────────────────────────────────────────────────
+
+# The `subagents` registry and its spawn/fork backends live in the HOST
+# composition: the registry is a process singleton whose cross-session queries
+# the api-proxy serves to the browser, and a provider name may only be
+# registered once. This preset contributes the delegation TOOLS, which resolve
+# that host registry.
+#
+# `workflows` is different — nothing outside an agent reads it — so every row
+# that reaches it shares one entry-local realm here, and a consumer left
+# outside would resolve a host registry this preset does not populate.
+- id: delegation
+  name: cordis:group
+  group: true
+  isolate:
+    workflowEngine: true
+  config:
+    - id: tool-subagent-control
+      name: '@deepseek-ai/dsh-tool-subagent-control'
+
+    - id: tool-subagent-list-agents
+      name: '@deepseek-ai/dsh-tool-subagent-control/list-agents'
+
+    - id: tool-subagent
+      name: '@deepseek-ai/dsh-tool-subagent'
+      config:
+        provider: spawn
+        toolName: subagent
+        modelSelectionSettings: true
+        backgroundMode: continuable
+
+    # Fork omits model selection so provider/model stay equal to the parent and
+    # the inherited history remains eligible for KV Cache reuse. This preset
+    # keeps fork continuable and accepts its child-scoped `report` additions invalidating
+    # that prefix; issue #2124 tracks cache-preserving continuable fork.
+    - id: tool-subagent-fork
+      name: '@deepseek-ai/dsh-tool-subagent'
+      config:
+        provider: fork
+        toolName: subagent_fork
+        backgroundMode: continuable
+
+    # Production dsh does not install these optional providers. Install the
+    # matching Bundle in this Profile and restart the Host, then copy this
+    # preset and remove `disabled` from the matching tool row. Host availability
+    # alone grants no tool.
+    - id: tool-subagent-codex
+      name: '@deepseek-ai/dsh-tool-subagent'
+      disabled: true
+      config:
+        provider: codex
+        toolName: subagent_codex
+        backgroundMode: one-shot
+        maxDepth: provider-managed
+
+    - id: tool-subagent-claude-code
+      name: '@deepseek-ai/dsh-tool-subagent'
+      disabled: true
+      config:
+        provider: claude-code
+        toolName: subagent_claude_code
+        backgroundMode: one-shot
+        maxDepth: provider-managed
+
+    - id: workflow-worker-thread
+      name: '@deepseek-ai/dsh-workflow-worker-thread'
+      config:
+        provider: spawn
+
+    - id: tool-workflow
+      name: '@deepseek-ai/dsh-tool-workflow'
+
+    - id: tool-ralph
+      name: '@deepseek-ai/dsh-tool-ralph'
+      config:
+        subagentProvider: spawn
+        maxRounds: 64
+
+# ── remaining model-facing rows ─────────────────────────────────────────────
+
+- id: tool-ask-user
+  name: '@deepseek-ai/dsh-tool-ask-user'
+
+- id: tool-todo
+  name: '@deepseek-ai/dsh-tool-todo'
+  config:
+    allowParallelInProgress: true
+
+# The `web` service and its search provider stay in the host composition; only
+# the model-facing tool is per-session.
+- id: tool-web
+  name: '@deepseek-ai/dsh-tool-web'
+  config:
+    fetch: true
+    searchTimeoutMs: 60000
+
+# ── presentation ────────────────────────────────────────────────────────────
+
+# PTC mode for this agent alone. The row waits for the host's `codeRuntime`
+# rather than assuming it: a deployment that composes no TypeScript runtime
+# fails this preset at mount, naming this id, instead of at the first request.
+- id: tool-presentation
+  name: '@deepseek-ai/dsh-agent-tool-presentation'
+  config:
+    mode: ptc

package/assets/code-gitbash/agent.cordis.yml

@@ -1,3 +1,7 @@
+# ERA: dsh <= 0.1.1 text (built-in preset id `code`, presented with mode: code).
+# The dsh >= 0.1.2 era text (built-in id `ptc`) lives in agent.cordis.ptc.yml; the
+# plugin probes the roster at materialization and writes the picked text as the
+# preset's own agent.cordis.yml.
 # The `code` agent preset: the standard coding agent, presented as Code Mode.
 #
 # Everything in `standard` is here unchanged. What is added is the `tool-presentation`

package/assets/code-gitbash/preset.yml

@@ -1,3 +1,3 @@
 name: PTC 模式 · Git Bash
-description: 具备标准模式的全部能力，并通过 Code Mode SDK 呈现工具，让模型用一个 TypeScript 程序组合多步操作。（Shell 使用 Git Bash）
+description: 具备标准模式的全部能力，并通过 PTC 模式 SDK 呈现工具，让模型用一个 TypeScript 程序组合多步操作。（Shell 使用 Git Bash）
 order: 2

package/assets/cordis-gitbash/agent.cordis.ptc.yml

@@ -0,0 +1,273 @@
+# ERA: dsh >= 0.1.2 text (renamed built-in rows, incl. the new `command-goal`).
+# The dsh <= 0.1.1 era text lives in agent.cordis.yml; the plugin probes the
+# roster at materialization and writes the picked text as the preset's own
+# agent.cordis.yml.
+# The `cordis` agent preset: the standard coding agent, plus the ability to
+# read and write the runtime it is running in.
+#
+# It exists so a person can ask an agent to author another agent. Everything in
+# `standard` is here unchanged; what is added is the self-referential Cordis
+# toolset, a skill that teaches composition authoring, and a persona that says
+# which of the two planes an edit belongs to.
+#
+# TRUST: `cordis_mount` evaluates model-written JavaScript against the live
+# runtime, and a composition this agent writes becomes a preset other sessions
+# mount. Treat a session on this preset as shell access — the toolset's own
+# documentation makes the same statement.
+
+
+# The preset's own persona, shadowing the deployment default for this agent.
+# `{{model}}` and `{{cwd}}` resolve from the agent's own route and workspace.
+- id: persona
+  name: '@deepseek-ai/dsh-persona'
+  config:
+    text: |-
+      You are a coding agent powered by the {{model}} model, running on the DeepSeek Harness. Your working directory is {{cwd}}.
+
+      You can read and modify the harness you run on. Its composition is Cordis: every capability is a plugin row in a `cordis.yml`, and an agent preset is one such file mounted for a single session.
+
+      Two planes decide where an edit belongs. The HOST composition holds the registries and anything shared across sessions — persistence, the sandbox and approval stack, the model route, the subagent registry and its backends. An AGENT PRESET holds what one session contributes to those registries: its tools, its persona, its prompt sections. A row that publishes a service belongs in the host composition, or inside an `isolate` realm if the preset genuinely owns that service and nothing outside one agent reads it.
+
+      Presets you author live one directory per preset under `${DSH_HOME:-$HOME/.dsh}/.agent-presets/<id>/`; the roster reports each preset's real path, so take the one you edit from there. NEVER edit or delete the shipped preset install (the `agent-presets` directory beside the deployment's own config): it belongs to the deployment, an upgrade overwrites it, and corrupting the `cordis` preset would disable this very mode. To change what a shipped preset does, copy its composition into a new preset directory and edit the copy.
+
+      Load the `editing-cordis-compositions` skill before writing or changing a composition.
+
+- id: agent-instructions
+  name: '@deepseek-ai/dsh-agent-instructions'
+  config:
+    maxBytes: 65536
+
+# ── shell ───────────────────────────────────────────────────────────────────
+
+# `shell-env` stays in the HOST composition: `apps/cli/src/web.ts` injects it to
+# publish `DSH_WEB_URL`/`DSH_WEB_MODE`, and a host row that injects a service is
+# the criterion for host-plane ownership — injection resolves before any session
+# exists, so there is no agent to key by. Behind a preset realm those variables
+# never reached the model's shell at all. Both shell tools consume the host
+# registry from here; the host executor is the Git Bash provider (`dsh-gitbash-shell`)
+# host-plane too.
+- id: tool-bash
+  name: '@deepseek-ai/dsh-tool-bash'
+  # Git Bash on Windows: the host executor is dsh-gitbash-shell (see $DSH_HOME/cordis.patch.yml).
+  disabled: false
+
+- id: tool-pwsh
+  name: '@deepseek-ai/dsh-tool-pwsh'
+  disabled: true
+
+# ── filesystem ──────────────────────────────────────────────────────────────
+
+# Both register into the host `tools` registry and provide nothing, so
+# they need no realm. The `fs` service and its policy stay in the host.
+- id: tool-fs
+  name: '@deepseek-ai/dsh-tool-fs'
+
+- id: tool-fs-search
+  name: '@deepseek-ai/dsh-tool-fs-search'
+  config:
+    sampleOverCapGlobResults: false
+
+# ── background jobs ────────────────────────────────────────────────────────
+
+# Only the model-facing controls. The task REGISTRY stays on the host plane:
+# its producers sit outside any realm this file could put it in — `tool-bash`
+# above resolves it with `ctx.get`, and an entry-local realm here is invisible
+# to every sibling row, so `run_in_background` would answer "background jobs
+# unavailable" while these controls sat in the catalog. The registry is keyed by
+# owning agent anyway, so one host instance serves every session. What a preset
+# chooses is whether its agent can collect and stop background work at all.
+- id: tool-jobs
+  name: '@deepseek-ai/dsh-tool-jobs'
+
+# ── goals ───────────────────────────────────────────────────────────────────
+
+# The goal service and session driver stay on the host plane, where the Gateway
+# can resolve them. The human command and model-facing tool register into this
+# preset's scoped layers.
+- id: command-goal
+  name: '@deepseek-ai/dsh-command-goal'
+
+- id: tool-goal
+  name: '@deepseek-ai/dsh-tool-goal'
+
+# ── plan mode ───────────────────────────────────────────────────────────────
+
+# Plan state is per-agent by nature, so an entry-local realm is not a
+# workaround here — it is the correct lifetime.
+- id: planning
+  name: cordis:group
+  group: true
+  isolate:
+    planMode: true
+  config:
+    - id: plan-mode
+      name: '@deepseek-ai/dsh-plan-mode'
+      config:
+        section: |
+              You are in plan mode. Stay in plan mode until exit_plan_mode succeeds or the user switches the session mode. Imperative language to implement changes means plan the implementation, not execute it. A user's conversational agreement — including an answer confirming something you asked — approves nothing and does not end plan mode; fold the confirmed decision into the plan and submit it through exit_plan_mode.
+
+              Explore first. Use non-mutating reads, searches, static analysis, and checks to ground the plan in the actual repository. Do not edit or write files, change configuration, run formatters or code generation that rewrites tracked files, commit, or otherwise carry out the plan. Prefer existing functions and patterns over new machinery.
+
+              The tool catalog stays the same across modes for request-cache stability. These plan-mode rules override any later tool description or guidance that suggests using mutation tools; those tools remain listed to keep the tool catalog unchanged. Do not use todo_write to track this planning phase: it tracks implementation after an approved plan, while the plan itself belongs in exit_plan_mode.
+
+              Resolve discoverable facts by inspection. Use ask_user_question only for user-owned choices or material ambiguity that inspection cannot answer. Do not ask the user where code lives or how current behavior works when you can find out.
+
+              Make the plan decision-complete: state the goal and success criteria; group implementation changes by subsystem; identify public API, schema, and data-flow changes; cover edge cases, failure modes, tests, acceptance criteria, and explicit assumptions. Keep it concise enough to review but detailed enough that another engineer can implement it without making design decisions.
+
+              When ready, call exit_plan_mode with the complete plan markdown, starting with a # title. Make exit_plan_mode the only and final tool call in that assistant response: it presents the plan for approval, and implementation begins only in a later step after approval. Do not paste the final plan as a plain reply or ask "should I proceed?" through prose or ask_user_question. If review rejects it, incorporate the feedback and present again. If the review channel is unavailable or aborted, stay in plan mode and ask the user to switch modes manually; do not proceed with implementation.
+
+# ── compaction ──────────────────────────────────────────────────────────────
+
+# `compaction-basic` reads `toolResultPrune` through `ctx.get`, so the pruner must
+# share this realm rather than sit outside it.
+#
+# `tokenMeter` is deliberately NOT in this realm: the meter stays on the HOST
+# plane, and the rows here resolve that one instance. It takes no configuration,
+# keys every fold by Session, and owns the context-meter projection units the
+# browser reads for every session — behind a realm those units would come and go
+# with whichever presets happen to be mounted. What a preset chooses is whether
+# its agent compacts at all, which is `compaction-basic` below.
+- id: compaction
+  name: cordis:group
+  group: true
+  isolate:
+    compaction: true
+    toolResultPruner: true
+  config:
+    - id: compaction-basic
+      name: '@deepseek-ai/dsh-compaction-basic'
+
+    - id: command-compact
+      name: '@deepseek-ai/dsh-command-compact'
+
+    - id: tool-result-pruner
+      name: '@deepseek-ai/dsh-compaction-tool-result-pruner'
+      config:
+        thresholdChars: 8192
+        headChars: 4096
+        tailChars: 1024
+
+# ── delegation and workflows ────────────────────────────────────────────────
+
+# The `subagents` registry and its spawn/fork backends live in the HOST
+# composition: the registry is a process singleton whose cross-session queries
+# the api-proxy serves to the browser, and a provider name may only be
+# registered once. This preset contributes the delegation TOOLS, which resolve
+# that host registry.
+#
+# `workflows` is different — nothing outside an agent reads it — so every row
+# that reaches it shares one entry-local realm here, and a consumer left
+# outside would resolve a host registry this preset does not populate.
+#
+# `tool-subagent-report` is host-plane for the same reason as the registry,
+# not because a preset may not want it: it registers a CONTINUABLE SETUP on
+# that singleton rather than a tool this agent calls, and the setup list is
+# not scope-aware — one copy per mounted preset means every child gets
+# `report` registered once per live session, which throws on the second.
+- id: delegation
+  name: cordis:group
+  group: true
+  isolate:
+    workflowEngine: true
+  config:
+    - id: tool-subagent-control
+      name: '@deepseek-ai/dsh-tool-subagent-control'
+
+    - id: tool-subagent-list-agents
+      name: '@deepseek-ai/dsh-tool-subagent-control/list-agents'
+
+    - id: tool-subagent
+      name: '@deepseek-ai/dsh-tool-subagent'
+      config:
+        provider: spawn
+        toolName: subagent
+        modelSelectionSettings: true
+        backgroundMode: continuable
+
+    # Fork omits model selection so provider/model stay equal to the parent and
+    # the inherited history remains eligible for KV Cache reuse. This preset
+    # keeps fork continuable and accepts its child-scoped `report` additions invalidating
+    # that prefix; issue #2124 tracks cache-preserving continuable fork.
+    - id: tool-subagent-fork
+      name: '@deepseek-ai/dsh-tool-subagent'
+      config:
+        provider: fork
+        toolName: subagent_fork
+        backgroundMode: continuable
+
+    # Production dsh does not install these optional providers. Install the
+    # matching Bundle in this Profile and restart the Host, then copy this
+    # preset and remove `disabled` from the matching tool row. Host availability
+    # alone grants no tool.
+    - id: tool-subagent-codex
+      name: '@deepseek-ai/dsh-tool-subagent'
+      disabled: true
+      config:
+        provider: codex
+        toolName: subagent_codex
+        backgroundMode: one-shot
+        maxDepth: provider-managed
+
+    - id: tool-subagent-claude-code
+      name: '@deepseek-ai/dsh-tool-subagent'
+      disabled: true
+      config:
+        provider: claude-code
+        toolName: subagent_claude_code
+        backgroundMode: one-shot
+        maxDepth: provider-managed
+
+    - id: workflow-worker-thread
+      name: '@deepseek-ai/dsh-workflow-worker-thread'
+      config:
+        provider: spawn
+
+    - id: tool-workflow
+      name: '@deepseek-ai/dsh-tool-workflow'
+
+    - id: tool-ralph
+      name: '@deepseek-ai/dsh-tool-ralph'
+      config:
+        subagentProvider: spawn
+        maxRounds: 64
+
+# ── remaining model-facing rows ─────────────────────────────────────────────
+
+- id: tool-ask-user
+  name: '@deepseek-ai/dsh-tool-ask-user'
+
+- id: tool-todo
+  name: '@deepseek-ai/dsh-tool-todo'
+  config:
+    allowParallelInProgress: true
+
+# The `web` service and its search provider stay in the host composition; only
+# the model-facing tool is per-session.
+- id: tool-web
+  name: '@deepseek-ai/dsh-tool-web'
+  config:
+    fetch: true
+    searchTimeoutMs: 60000
+
+# ── self-modification ───────────────────────────────────────────────────────
+
+# Read the live runtime, mount a temporary plugin, unmount it. The toolset is a
+# trust boundary, not a sandbox — see this file's header.
+- id: tool-cordis
+  name: '@deepseek-ai/dsh-tool-cordis'
+
+# The composition-authoring skill travels with this preset rather than living
+# in the user's skill root: it documents THIS deployment's two planes, and a
+# preset is the unit that gets copied and edited. `baseUrl` is the preset's
+# own directory, so the root resolves wherever the preset is installed.
+# Both rows register into THIS preset's layer of the host skill registry, so
+# they need no realm; the agent's merged catalog also carries whatever the
+# deployment registered globally (repository plugins).
+- id: skill-filesystem
+  name: '@deepseek-ai/dsh-skill-filesystem'
+  config:
+    customSkillDirs:
+      - !!js "process.getBuiltinModule('node:url').fileURLToPath(new URL('skills/', baseUrl))"
+
+- id: tool-skill
+  name: '@deepseek-ai/dsh-tool-skill'

package/assets/cordis-gitbash/agent.cordis.yml

@@ -1,3 +1,6 @@
+# ERA: dsh <= 0.1.1 text (pre-rename built-in rows). The dsh >= 0.1.2 era text
+# lives in agent.cordis.ptc.yml; the plugin probes the roster at materialization
+# and writes the picked text as the preset's own agent.cordis.yml.
 # The `cordis` agent preset: the standard coding agent, plus the ability to
 # read and write the runtime it is running in.
 #

package/assets/minimal-gitbash/agent.cordis.yml

@@ -1,3 +1,7 @@
+# ERA: one text serves both dsh eras — the built-in `minimal` preset did not
+# change across the 0.1.1 → 0.1.2 code→ptc rename, so no agent.cordis.ptc.yml
+# twin exists here. The plugin probes the roster and writes this file as the
+# preset's own agent.cordis.yml either way.
 # The `minimal` agent preset: a fixed-prompt, two-tool coding-agent composition.
 #
 # The persona is the complete system prompt, so global identity, Web orientation,

package/assets/standard-gitbash/agent.cordis.ptc.yml

@@ -0,0 +1,262 @@
+# ERA: dsh >= 0.1.2 text (renamed built-in rows, incl. the new `command-goal`).
+# The dsh <= 0.1.1 era text lives in agent.cordis.yml; the plugin probes the
+# roster at materialization and writes the picked text as the preset's own
+# agent.cordis.yml.
+# The `standard` agent preset: the full coding agent, mounted once per process.
+#
+# This file is an AGENT-PLANE composition. The roster mounts it ONCE under a
+# standing scope; every session naming it joins by scope parentage, so the
+# tools and prompt sections registered here cover each joined agent while a
+# session's own state stays keyed per Session/Agent inside the plugins. The
+# host composition (`base.cordis.yml` + `web.cordis.yml`) keeps everything a
+# preset must not own: the registries themselves, the sandbox and approval
+# stack, persistence, and the model route.
+#
+# A service row here MUST sit inside a group carrying an `isolate` realm.
+# Without one it publishes into the root realm, where it is process-global —
+# another preset publishing the same name collides, and a host reader would
+# resolve one preset's instance for every session; `dsh-agent-presets` rejects
+# that at mount. `true` means an entry-local realm: this standing mount's own
+# private instance, apart from every other preset's. (A shared label does NOT
+# pool instances — `provide()` throws on the second registration under the
+# same realm symbol; labels join REALMS, and are not what this file needs.)
+
+# ── identity ────────────────────────────────────────────────────────────────
+
+# The preset's own persona, shadowing the deployment default for this agent.
+# `{{model}}` and `{{cwd}}` resolve from the agent's own route and workspace.
+- id: persona
+  name: '@deepseek-ai/dsh-persona'
+  config:
+    text: >-
+      You are a coding agent powered by the {{model}} model. Your working directory is {{cwd}}.
+
+- id: agent-instructions
+  name: '@deepseek-ai/dsh-agent-instructions'
+  config:
+    maxBytes: 65536
+
+# ── shell ───────────────────────────────────────────────────────────────────
+
+# `shell-env` stays in the HOST composition: `apps/cli/src/web.ts` injects it to
+# publish `DSH_WEB_URL`/`DSH_WEB_MODE`, and a host row that injects a service is
+# the criterion for host-plane ownership — injection resolves before any session
+# exists, so there is no agent to key by. Behind a preset realm those variables
+# never reached the model's shell at all. Both shell tools consume the host
+# registry from here; the host executor is the Git Bash provider (`dsh-gitbash-shell`)
+# host-plane too.
+- id: tool-bash
+  name: '@deepseek-ai/dsh-tool-bash'
+  # Git Bash on Windows: the host executor is dsh-gitbash-shell (see $DSH_HOME/cordis.patch.yml).
+  disabled: false
+
+- id: tool-pwsh
+  name: '@deepseek-ai/dsh-tool-pwsh'
+  disabled: true
+
+# ── filesystem ──────────────────────────────────────────────────────────────
+
+# Both register into the host `tools` registry and provide nothing, so
+# they need no realm. The `fs` service and its policy stay in the host.
+- id: tool-fs
+  name: '@deepseek-ai/dsh-tool-fs'
+
+- id: tool-fs-search
+  name: '@deepseek-ai/dsh-tool-fs-search'
+  config:
+    sampleOverCapGlobResults: false
+
+# ── background jobs ────────────────────────────────────────────────────────
+
+# Only the model-facing controls. The task REGISTRY stays on the host plane:
+# its producers sit outside any realm this file could put it in — `tool-bash`
+# above resolves it with `ctx.get`, and an entry-local realm here is invisible
+# to every sibling row, so `run_in_background` would answer "background jobs
+# unavailable" while these controls sat in the catalog. The registry is keyed by
+# owning agent anyway, so one host instance serves every session. What a preset
+# chooses is whether its agent can collect and stop background work at all.
+- id: tool-jobs
+  name: '@deepseek-ai/dsh-tool-jobs'
+
+# ── skills ──────────────────────────────────────────────────────────────────
+
+# The skill REGISTRY lives in the host composition and is layered per scope:
+# these rows register into THIS preset's layer of it, so they need no realm.
+# `skill-filesystem` contributes local-root discovery for agents on this preset, and
+# `tool-skill` gives them the catalog and loader; the merged catalog also
+# carries whatever the deployment registered globally (repository plugins).
+- id: skill-filesystem
+  name: '@deepseek-ai/dsh-skill-filesystem'
+
+- id: tool-skill
+  name: '@deepseek-ai/dsh-tool-skill'
+
+# ── goals ───────────────────────────────────────────────────────────────────
+
+# The goal service and session driver stay on the host plane, where the Gateway
+# can resolve them. The human command and model-facing tool register into this
+# preset's scoped layers.
+- id: command-goal
+  name: '@deepseek-ai/dsh-command-goal'
+
+- id: tool-goal
+  name: '@deepseek-ai/dsh-tool-goal'
+
+# ── plan mode ───────────────────────────────────────────────────────────────
+
+# Plan state is per-agent by nature, so an entry-local realm is not a
+# workaround here — it is the correct lifetime.
+- id: planning
+  name: cordis:group
+  group: true
+  isolate:
+    planMode: true
+  config:
+    - id: plan-mode
+      name: '@deepseek-ai/dsh-plan-mode'
+      config:
+        section: |
+              You are in plan mode. Stay in plan mode until exit_plan_mode succeeds or the user switches the session mode. Imperative language to implement changes means plan the implementation, not execute it. A user's conversational agreement — including an answer confirming something you asked — approves nothing and does not end plan mode; fold the confirmed decision into the plan and submit it through exit_plan_mode.
+
+              Explore first. Use non-mutating reads, searches, static analysis, and checks to ground the plan in the actual repository. Do not edit or write files, change configuration, run formatters or code generation that rewrites tracked files, commit, or otherwise carry out the plan. Prefer existing functions and patterns over new machinery.
+
+              The tool catalog stays the same across modes for request-cache stability. These plan-mode rules override any later tool description or guidance that suggests using mutation tools; those tools remain listed to keep the tool catalog unchanged. Do not use todo_write to track this planning phase: it tracks implementation after an approved plan, while the plan itself belongs in exit_plan_mode.
+
+              Resolve discoverable facts by inspection. Use ask_user_question only for user-owned choices or material ambiguity that inspection cannot answer. Do not ask the user where code lives or how current behavior works when you can find out.
+
+              Make the plan decision-complete: state the goal and success criteria; group implementation changes by subsystem; identify public API, schema, and data-flow changes; cover edge cases, failure modes, tests, acceptance criteria, and explicit assumptions. Keep it concise enough to review but detailed enough that another engineer can implement it without making design decisions.
+
+              When ready, call exit_plan_mode with the complete plan markdown, starting with a # title. Make exit_plan_mode the only and final tool call in that assistant response: it presents the plan for approval, and implementation begins only in a later step after approval. Do not paste the final plan as a plain reply or ask "should I proceed?" through prose or ask_user_question. If review rejects it, incorporate the feedback and present again. If the review channel is unavailable or aborted, stay in plan mode and ask the user to switch modes manually; do not proceed with implementation.
+
+# ── compaction ──────────────────────────────────────────────────────────────
+
+# `compaction-basic` reads `toolResultPrune` through `ctx.get`, so the pruner must
+# share this realm rather than sit outside it.
+#
+# `tokenMeter` is deliberately NOT in this realm: the meter stays on the HOST
+# plane, and the rows here resolve that one instance. It takes no configuration,
+# keys every fold by Session, and owns the context-meter projection units the
+# browser reads for every session — behind a realm those units would come and go
+# with whichever presets happen to be mounted. What a preset chooses is whether
+# its agent compacts at all, which is `compaction-basic` below.
+- id: compaction
+  name: cordis:group
+  group: true
+  isolate:
+    compaction: true
+    toolResultPruner: true
+  config:
+    - id: compaction-basic
+      name: '@deepseek-ai/dsh-compaction-basic'
+
+    - id: command-compact
+      name: '@deepseek-ai/dsh-command-compact'
+
+    - id: tool-result-pruner
+      name: '@deepseek-ai/dsh-compaction-tool-result-pruner'
+      config:
+        thresholdChars: 8192
+        headChars: 4096
+        tailChars: 1024
+
+# ── delegation and workflows ────────────────────────────────────────────────
+
+# The `subagents` registry and its spawn/fork backends live in the HOST
+# composition: the registry is a process singleton whose cross-session queries
+# the api-proxy serves to the browser, and a provider name may only be
+# registered once. This preset contributes the delegation TOOLS, which resolve
+# that host registry.
+#
+# `workflows` is different — nothing outside an agent reads it — so every row
+# that reaches it shares one entry-local realm here, and a consumer left
+# outside would resolve a host registry this preset does not populate.
+#
+# `tool-subagent-report` is host-plane for the same reason as the registry,
+# not because a preset may not want it: it registers a CONTINUABLE SETUP on
+# that singleton rather than a tool this agent calls, and the setup list is
+# not scope-aware — one copy per mounted preset means every child gets
+# `report` registered once per live session, which throws on the second.
+- id: delegation
+  name: cordis:group
+  group: true
+  isolate:
+    workflowEngine: true
+  config:
+    - id: tool-subagent-control
+      name: '@deepseek-ai/dsh-tool-subagent-control'
+
+    - id: tool-subagent-list-agents
+      name: '@deepseek-ai/dsh-tool-subagent-control/list-agents'
+
+    - id: tool-subagent
+      name: '@deepseek-ai/dsh-tool-subagent'
+      config:
+        provider: spawn
+        toolName: subagent
+        modelSelectionSettings: true
+        backgroundMode: continuable
+
+    # Fork omits model selection so provider/model stay equal to the parent and
+    # the inherited history remains eligible for KV Cache reuse. This preset
+    # keeps fork continuable and accepts its child-scoped `report` additions invalidating
+    # that prefix; issue #2124 tracks cache-preserving continuable fork.
+    - id: tool-subagent-fork
+      name: '@deepseek-ai/dsh-tool-subagent'
+      config:
+        provider: fork
+        toolName: subagent_fork
+        backgroundMode: continuable
+
+    # Production dsh does not install these optional providers. Install the
+    # matching Bundle in this Profile and restart the Host, then copy this
+    # preset and remove `disabled` from the matching tool row. Host availability
+    # alone grants no tool.
+    - id: tool-subagent-codex
+      name: '@deepseek-ai/dsh-tool-subagent'
+      disabled: true
+      config:
+        provider: codex
+        toolName: subagent_codex
+        backgroundMode: one-shot
+        maxDepth: provider-managed
+
+    - id: tool-subagent-claude-code
+      name: '@deepseek-ai/dsh-tool-subagent'
+      disabled: true
+      config:
+        provider: claude-code
+        toolName: subagent_claude_code
+        backgroundMode: one-shot
+        maxDepth: provider-managed
+
+    - id: workflow-worker-thread
+      name: '@deepseek-ai/dsh-workflow-worker-thread'
+      config:
+        provider: spawn
+
+    - id: tool-workflow
+      name: '@deepseek-ai/dsh-tool-workflow'
+
+    - id: tool-ralph
+      name: '@deepseek-ai/dsh-tool-ralph'
+      config:
+        subagentProvider: spawn
+        maxRounds: 64
+
+# ── remaining model-facing rows ─────────────────────────────────────────────
+
+- id: tool-ask-user
+  name: '@deepseek-ai/dsh-tool-ask-user'
+
+- id: tool-todo
+  name: '@deepseek-ai/dsh-tool-todo'
+  config:
+    allowParallelInProgress: true
+
+# The `web` service and its search provider stay in the host composition; only
+# the model-facing tool is per-session.
+- id: tool-web
+  name: '@deepseek-ai/dsh-tool-web'
+  config:
+    fetch: true
+    searchTimeoutMs: 60000

package/assets/standard-gitbash/agent.cordis.yml

@@ -1,3 +1,6 @@
+# ERA: dsh <= 0.1.1 text (pre-rename built-in rows). The dsh >= 0.1.2 era text
+# lives in agent.cordis.ptc.yml; the plugin probes the roster at materialization
+# and writes the picked text as the preset's own agent.cordis.yml.
 # The `standard` agent preset: the full coding agent, mounted once per process.
 #
 # This file is an AGENT-PLANE composition. The roster mounts it ONCE under a

package/dsh.plugin.json

@@ -1,8 +1,8 @@
 {
   "id": "dsh-external/dsh-gitbash-shell",
-  "version": "0.5.4",
+  "version": "0.6.0",
   "main": "./src/index.js",
-  "description": "Windows 全模式 Git Bash:用 Git for Windows bash 替换 PowerShell 执行器,并物化 standard/minimal/code/cordis 的 Git Bash 变体 preset。联动 dsh-ptc-cordis-preset:两者同装时 PTC 创造模式自动物化为 Git Bash 版。",
+  "description": "Windows 全模式 Git Bash:用 Git for Windows bash 替换 PowerShell 执行器,并物化 standard/minimal/code/cordis 的 Git Bash 变体 preset。双 era 适配 dsh <= 0.1.1(内置 id 'code')与 >= 0.1.2(改名 'ptc'),启动探测自动选择,preset id 永不变。联动:dsh-ptc-cordis-preset(两者同装时 PTC 创造模式自动物化为 Git Bash 版)+ dsh-better-sidebar(终端 UI 标签与模型 terminal_* 工具自动切到 Git Bash)。",
   "engines": {
     "dsh": ">=0.0.1"
   },

package/package.json

@@ -1,7 +1,7 @@
 {
   "name": "dsh-gitbash-shell",
-  "version": "0.5.4",
-  "description": "DSH plugin: run every agent shell command through Git for Windows bash on Windows instead of PowerShell. Replaces the pwsh executor with a Git Bash ctx.shell provider and materializes Git Bash variants of the standard/minimal/code/cordis agent presets into your user preset root at startup.",
+  "version": "0.6.0",
+  "description": "DSH plugin: run every agent shell command through Git for Windows bash on Windows instead of PowerShell. Replaces the pwsh executor with a Git Bash ctx.shell provider and materializes Git Bash variants of the standard/minimal/code/cordis agent presets into your user preset root at startup. On Windows it also adopts dsh-better-sidebar's terminal shell (UI tabs and the model-facing terminal_* tools) through the sidebar's official settings seam, and pairs with dsh-ptc-cordis-preset so the PTC Creation-mode preset materializes as its Git Bash variant. Ships both composition eras for dsh <= 0.1.1 (built-in preset id 'code') and dsh >= 0.1.2 (renamed id 'ptc') — the era is probed per boot and re-materialized automatically after a dsh upgrade, in either upgrade order; preset ids never change so pinned sessions stay valid.",
   "type": "module",
   "repository": {
     "type": "git",

package/src/index.js

@@ -28,6 +28,16 @@
  * directory that still exists means reload/update/restart — keep the
  * presets; a vanished package.json means uninstall — remove each preset the
  * user never modified.
+ *
+ * ERA SPLIT (v0.6.0): dsh renamed the built-in `code` preset to `ptc` in
+ * 0.1.2 with no compatibility alias (tool-presentation `mode` value, plus new
+ * built-in rows), so each variant ships BOTH committed era texts where the
+ * built-in changed (minimal did not — one text serves both). The era is
+ * probed per boot from the roster (`ptc` → dsh >= 0.1.2, else `code`),
+ * recorded in the marker (`base`), and a flipped detection re-materializes
+ * on the next startup — either upgrade order converges, and preset IDS never
+ * change (sessions are pinned to them; `code-gitbash` keeps its historical
+ * id even though its dsh >= 0.1.2 text presents through the `ptc` built-in).
  */
 
 import { createHash } from 'node:crypto'
@@ -128,10 +138,56 @@ function skillsHashes(source) {
   return out
 }
 
+// ── built-in era detection (dsh 0.1.2 renamed the `code` preset to `ptc`) ──
+
+/**
+ * Which built-in PTC preset does the installed dsh ship? dsh 0.1.2 renamed
+ * the preset id `code` to `ptc` with no compatibility alias, so the variant
+ * compositions are era-specific where the built-in changed. The roster probe
+ * is the version-agnostic signal. Pure companion of `detectBase`; 'ptc' wins
+ * if both ids somehow exist, and an unknown roster conservatively maps to
+ * 'code' — the era text the stable release accepted.
+ */
+function baseForRoster(ids) {
+  const set = new Set(ids)
+  if (set.has('ptc')) return 'ptc'
+  return 'code'
+}
+
+/** Async probe against the live roster; never throws. */
+async function detectBase(agentPresets) {
+  try {
+    const list = await agentPresets.list()
+    return baseForRoster((Array.isArray(list) ? list : []).map((p) => p && p.id))
+  } catch (error) {
+    console.log(`${TAG} built-in preset probe failed (${error?.message ?? error}) — assuming the 'code' era`)
+    return 'code'
+  }
+}
+
+/** Era suffix for committed composition assets: ptc-era files carry `.ptc`. */
+function eraSuffix(base) {
+  return base === 'ptc' ? '.ptc' : ''
+}
+
+/**
+ * Pick the committed composition asset inside one variant directory (pure).
+ * The ptc-era twin is preferred when the roster says `ptc`; variants without
+ * a twin (minimal — its built-in did not change) fall back to the plain base
+ * file. Every candidate is a committed file — no runtime text synthesis.
+ */
+function pickComposition(base, available) {
+  const era = eraSuffix(base)
+  const candidates = [`agent.cordis${era}.yml`, 'agent.cordis.yml']
+  for (const file of candidates) if (available.includes(file)) return file
+  return 'agent.cordis.yml'
+}
+
 /** Startup decision for an existing unmodified tree: 'refresh' or 'idle'. */
-function syncDecision({ state, marker, version, sourceHashes }) {
+function syncDecision({ state, marker, version, sourceHashes, base = 'code' }) {
   if (state !== 'unmodified' || !marker) return 'refresh'
   if (marker.version !== version) return 'refresh'
+  if (marker.base !== base) return 'refresh'
   const recorded = {}
   for (const k of Object.keys(marker.files)) if (k.startsWith('skills/')) recorded[k] = marker.files[k]
   if (sourceHashes === null) return Object.keys(recorded).length === 0 ? 'idle' : 'refresh'
@@ -145,11 +201,13 @@ function syncDecision({ state, marker, version, sourceHashes }) {
 // ── materialization ─────────────────────────────────────────────────────────
 
 /** Write one preset directory from scratch. Returns 'ok' or 'no-skills-source'. */
-function materialize({ target, presetId, skillsSource, version }) {
+function materialize({ target, presetId, skillsSource, version, base = 'code' }) {
   rmSync(target, { recursive: true, force: true })
   mkdirSync(target, { recursive: true })
 
-  writeFileSync(join(target, 'agent.cordis.yml'), readFileSync(join(pkgDir, 'assets', presetId, 'agent.cordis.yml')))
+  const available = readdirSync(join(pkgDir, 'assets', presetId)).filter((f) => f.endsWith('.yml'))
+  const compositionFile = pickComposition(base, available)
+  writeFileSync(join(target, 'agent.cordis.yml'), readFileSync(join(pkgDir, 'assets', presetId, compositionFile)))
   writeFileSync(join(target, 'preset.yml'), readFileSync(join(pkgDir, 'assets', presetId, 'preset.yml')))
 
   let skills = 'none'
@@ -163,7 +221,7 @@ function materialize({ target, presetId, skillsSource, version }) {
     }
   }
 
-  const marker = { managedBy: MANAGED_BY, version, presetId, files: hashTree(target) }
+  const marker = { managedBy: MANAGED_BY, version, presetId, base, files: hashTree(target) }
   writeFileSync(join(target, MARKER_FILE), JSON.stringify(marker, null, 2) + '\n')
   return skills
 }
@@ -413,6 +471,7 @@ export async function apply(ctx, config = {}) {
   }
 
   const skillsSource = await findSkillsSource(ctx.agentPresets)
+  const base = await detectBase(ctx.agentPresets)
   const userRootPath = userRoot.path
   purgeOrphans(userRootPath, presetIds)
 
@@ -442,19 +501,19 @@ export async function apply(ctx, config = {}) {
 
     const sourceHashes = presetId === 'cordis-gitbash' ? skillsHashes(skillsSource) : null
     const marker = readMarker(target)
-    if (state === 'unmodified' && syncDecision({ state, marker, version, sourceHashes }) === 'idle') {
-      ctx.logger?.('gitbash-shell')?.debug?.( `preset '${presetId}' up to date (v${version}) — idle`)
+    if (state === 'unmodified' && syncDecision({ state, marker, version, sourceHashes, base }) === 'idle') {
+      ctx.logger?.('gitbash-shell')?.debug?.( `preset '${presetId}' up to date (v${version}, ${base}-era) — idle`)
       continue
     }
 
-    const skills = materialize({ target, presetId, skillsSource, version })
+    const skills = materialize({ target, presetId, skillsSource, version, base })
     const verb = state === 'absent' ? 'materialized' : 'refreshed'
     console.log(
-      `${TAG} ${verb} preset '${presetId}' into ${userRootPath} (v${version})` +
+      `${TAG} ${verb} preset '${presetId}' into ${userRootPath} (v${version}, ${base}-era composition)` +
         (skills === 'copied' ? " (skills copied from the installed 'cordis' preset)" : '')
     )
   }
 }
 
 // Test surface: pure helpers, no Cordis context required.
-export const _internal = { PRESET_IDS, MARKER_FILE, classify, materialize, cleanupOnDispose, firstUserRoot, hashTree, skillsHashes, syncDecision, installRegisterShim }
+export const _internal = { PRESET_IDS, MARKER_FILE, classify, materialize, cleanupOnDispose, firstUserRoot, hashTree, skillsHashes, syncDecision, installRegisterShim, baseForRoster, detectBase, pickComposition }