@blxzer/cursor-trellis 0.2.8 → 0.3.0

package/dist/templates/markdown/spec/guides/durable-learning-decision-guide.md.txt

@@ -1,4 +1,4 @@
-# Durable Learning Decision Guide
+# Durable Learning Decision Guide
 
 > **Purpose**: Decide whether a completed task should update `.trellis/spec/` or can archive with an explicit **no durable learning** line.
 
@@ -17,7 +17,7 @@ Pick **one** grep-friendly line before `archive --check`:
 | Insight already documented | `Learning artifact: .trellis/tasks/<task>/handoff.md` |
 | Retrospective captured | `Learning artifact: .trellis/workspace/...` or `retrospective.md: <path>` |
 
-Optional planning note (not sufficient alone): `Spec update needed: <reason>` — still record **spec update evidence** after `/trellis:update-spec`.
+Optional planning note (not sufficient alone): `Spec update needed: <reason>` — still record **spec update evidence** after `/cstl:update-spec`.
 
 ---
 
@@ -33,7 +33,7 @@ Optional planning note (not sufficient alone): `Spec update needed: <reason>`
 
 - New or changed CLI/API signature or gate behavior
 - Repeated parent review **changes** on the same class of mistake
-- Workflow bug that should not recur (after `/trellis:break-loop` analysis)
+- Workflow bug that should not recur (after `/cstl:break-loop` analysis)
 - Cross-layer contract clarification
 
 ### Learning artifact
@@ -60,9 +60,9 @@ python ./.trellis/scripts/task.py review-child <parent> <child> --check
 
 ## Skills
 
-- **`/trellis:update-spec`** — edit code-spec after reviewer confirmation
-- **`/trellis:break-loop`** — deep bug analysis; often leads to spec updates
-- **`/trellis:finish-work`** — archive after learning line is present in `verify.md`
+- **`/cstl:update-spec`** — edit code-spec after reviewer confirmation
+- **`/cstl:break-loop`** — deep bug analysis; often leads to spec updates
+- **`/cstl:finish-work`** — archive after learning line is present in `verify.md`
 
 ---
 

package/dist/templates/markdown/spec/guides/execution-strategy.md.txt

@@ -8,7 +8,7 @@
 
 | Field | Question |
 | --- | --- |
-| `execution_mode` | **Who** implements and checks? (`inline` main session, `worker` trellis-implement/check agents, `child-task` Child worker) |
+| `execution_mode` | **Who** implements and checks? (`inline` main session, `worker` cstl-implement/check agents, `child-task` Child worker) |
 | `isolation` | **Where** are files edited? (`main-worktree` vs `git-worktree` for Child) |
 
 Non-git repo roots do **not** force `inline` for code Full tasks; they only affect whether `git-worktree` / `prepare-child-worktree` is recommended.
@@ -40,4 +40,4 @@ Run during brainstorm / before finalizing the YAML block in `implement.md`. Reco
 
 ## Phase 2 alignment
 
-See `workflow.md` Phase 2.1 / 2.2 tables: spawn `trellis-implement` / `trellis-check` only when `execution_mode: worker`.
+See `workflow.md` Phase 2.1 / 2.2 tables: spawn `cstl-implement` / `cstl-check` only when `execution_mode: worker`.

package/dist/templates/markdown/spec/guides/index.md.txt

@@ -1,4 +1,4 @@
-# Thinking Guides
+# Thinking Guides
 
 > **Purpose**: Expand your thinking to catch things you might not have considered.
 
@@ -21,10 +21,11 @@ These guides help you **ask the right questions before coding**.
 
 | Guide | Purpose | When to Use |
 |-------|---------|-------------|
-| [Durable Learning Decision Guide](./durable-learning-decision-guide.md) | Choose no-learning vs spec update vs learning artifact before archive | Finishing a task, parent review, or after `/trellis:break-loop` |
+| [Durable Learning Decision Guide](./durable-learning-decision-guide.md) | Choose no-learning vs spec update vs learning artifact before archive | Finishing a task, parent review, or after `/cstl:break-loop` |
 | [Code Reuse Thinking Guide](./code-reuse-thinking-guide.md) | Identify patterns and reduce duplication | When you notice repeated patterns |
 | [Cross-Layer Thinking Guide](./cross-layer-thinking-guide.md) | Think through data flow across layers | Features spanning multiple layers |
 | [Retrieval daily guide](./retrieval-daily-guide.md) | rg / codegraph / fast-context / smart-search / pack / router | Planning, research, finish verification |
+| [Cursor semantic compliance](./cursor-semantic-compliance.md) | Native planned vs executed semantic (`SemanticSearch`, classify_tool_calls) | Retrieval eval, OC-15 compliance runs, telemetry interpretation |
 | [Cursor subagent policy](./cursor-subagent-policy.md) | Task tool scenes, Agent mode, `model_policy: cursor-configured`, Cursor++ | Parent/Child dispatch, research/implement/check, parallel execution |
 | [Execution strategy](./execution-strategy.md) | `suggest-execution-strategy`, rules JSON, inline vs worker vs child-task | Writing `implement.md` contract, Phase 2 dispatch |
 | [Cursor context injection guide](./cursor-context-injection-guide.md) | Which Cursor channels reach the model (sessionStart bug, .cursor/rules, AGENTS.md) | "Agent ignored instructions" on Cursor, adding always-visible rules |

package/dist/templates/markdown/spec/guides/retrieval-daily-guide.md.txt

@@ -78,6 +78,18 @@ When `smart-search doctor` is not ok or `run_smart_search.py` status is `not_con
 
 **CLI discovery (Cursor):** `TRELLIS_SMART_SEARCH_COMMAND` / `smart_search.command` → PATH `smart-search` → project `node_modules/.bin/smart-search` (when installed as a dependency). Agent entrypoint is always `./.trellis/scripts/run_smart_search.py`.
 
+**Research flags (smart-search ≥ 0.1.15):** pass through `run_smart_search.py` for `--intent deep-research`:
+
+| Flag | Values | Use |
+| --- | --- | --- |
+| `--locale-scope` | `cn`, `en`, `both` (CLI default) | Skip bilingual discovery when cost matters |
+| `--dry-run` | flag | Plan + routing preview only; no live providers |
+| `--progress` | flag | `[research]` stage logs to stderr |
+
+Research JSON may include `output_schema_version: 1` and structured `citations` (`id`, `source_type`, `verified`, `content_len`). Trellis manifests preserve those fields when present.
+
+**Cursor skill surface:** `smart-search-cli` is an internal workflow name — it is **not** installed under `.cursor/skills/` (commands-only policy). On Cursor, follow this guide + `run_smart_search.py`; bundled skill assets ship for Codex/Gemini via `.agents/skills/` only.
+
 ## Router vs execution
 
 `route_codebase_retrieval.py` returns **intent + route suggestions** and **`agentInstructions`** (numbered steps with Cursor-native tool names). **codebase-evidence** is **candidate** until confirmed by Read/Git/tests.

package/dist/templates/shared-hooks/index.d.ts

@@ -16,7 +16,7 @@ export type SharedHookPlatform = "claude" | "cursor" | "codex" | "gemini" | "qod
 /**
  * Which shared hooks each platform actually invokes. Single source of truth
  * for shared-hook distribution — both `writeSharedHooks` (runtime install)
- * and `collectSharedHooks` (`trellis update` diff) read from this table.
+ * and `collectSharedHooks` (`cstl update` diff) read from this table.
  *
  * Routing rules encoded here:
  * - `session-start.py` — shipped by every platform with a SessionStart

package/dist/templates/shared-hooks/index.js

@@ -16,7 +16,7 @@ function readTemplate(relativePath) {
 /**
  * Which shared hooks each platform actually invokes. Single source of truth
  * for shared-hook distribution — both `writeSharedHooks` (runtime install)
- * and `collectSharedHooks` (`trellis update` diff) read from this table.
+ * and `collectSharedHooks` (`cstl update` diff) read from this table.
  *
  * Routing rules encoded here:
  * - `session-start.py` — shipped by every platform with a SessionStart

package/dist/templates/shared-hooks/inject-subagent-context.py

@@ -28,9 +28,9 @@ if sys.platform.startswith("win"):
 
 DIR_WORKFLOW = ".trellis"
 
-AGENT_IMPLEMENT = "trellis-implement"
-AGENT_CHECK = "trellis-check"
-AGENT_RESEARCH = "trellis-research"
+AGENT_IMPLEMENT = "cstl-implement"
+AGENT_CHECK = "cstl-check"
+AGENT_RESEARCH = "cstl-research"
 AGENTS_ALL = (AGENT_IMPLEMENT, AGENT_CHECK, AGENT_RESEARCH)
 AGENTS_REQUIRE_TASK = (AGENT_IMPLEMENT, AGENT_CHECK)
 

package/dist/templates/shared-hooks/inject-workflow-state.py

@@ -1,9 +1,9 @@
 #!/usr/bin/env python3
 """Trellis per-turn breadcrumb hook (UserPromptSubmit / BeforeAgent equivalent).
 
-Runs on every user prompt. Resolves the selected task through Trellis'
-session-aware selected task resolver and emits a short <workflow-state>
-block reminding the main AI what task is selected and its expected flow.
+Runs on every user prompt. Resolves the selected task through Trellis'
+session-aware selected task resolver and emits a short <workflow-state>
+block reminding the main AI what task is selected and its expected flow.
 
 The emitted ``hookEventName`` field is platform-aware: most hosts expect
 ``UserPromptSubmit`` (Claude Code naming, also accepted by Cursor / Qoder /
@@ -58,11 +58,11 @@ if sys.platform.startswith("win"):
 from typing import Optional
 
 
-# Bootstrap notice for Codex while the session has no selected task. Codex does not
+# Bootstrap notice for Codex while the session has no selected task. Codex does not
 # get the full SessionStart overview; this short reminder points the main session
 # at the start skill once and leaves the per-turn state block compact.
 CODEX_NO_TASK_BOOTSTRAP_NOTICE = """<trellis-bootstrap>
-If you have not already loaded Trellis context this session, read the `trellis-start` skill once.
+If you have not already loaded Trellis context this session, read the `cstl-start` skill once.
 </trellis-bootstrap>"""
 
 
@@ -85,7 +85,7 @@ def find_trellis_root(start: Path) -> Optional[Path]:
 
 
 # ---------------------------------------------------------------------------
-# Selected task discovery
+# Selected task discovery
 # ---------------------------------------------------------------------------
 
 def _detect_platform(input_data: dict) -> str | None:
@@ -124,18 +124,18 @@ def _detect_platform(input_data: dict) -> str | None:
     return None
 
 
-def _resolve_selected_task(root: Path, input_data: dict):
-    scripts_dir = root / ".trellis" / "scripts"
-    if str(scripts_dir) not in sys.path:
-        sys.path.insert(0, str(scripts_dir))
-    from common.active_task import resolve_selected_task  # type: ignore[import-not-found]
-
-    return resolve_selected_task(root, input_data, platform=_detect_platform(input_data))
-
-
-def get_selected_task(root: Path, input_data: dict) -> Optional[tuple[str, str, str]]:
-    """Return (task_id, status, source) from the selected task."""
-    active = _resolve_selected_task(root, input_data)
+def _resolve_selected_task(root: Path, input_data: dict):
+    scripts_dir = root / ".trellis" / "scripts"
+    if str(scripts_dir) not in sys.path:
+        sys.path.insert(0, str(scripts_dir))
+    from common.active_task import resolve_selected_task  # type: ignore[import-not-found]
+
+    return resolve_selected_task(root, input_data, platform=_detect_platform(input_data))
+
+
+def get_selected_task(root: Path, input_data: dict) -> Optional[tuple[str, str, str]]:
+    """Return (task_id, status, source) from the selected task."""
+    active = _resolve_selected_task(root, input_data)
     if not active.task_path:
         return None
 
@@ -292,11 +292,11 @@ def build_breadcrumb(
         body = templates.get(status)
     if body is None:
         body = "Refer to workflow.md for current step."
-    header = (
-        f"Selected task: none\nStatus: {status}"
-        if task_id is None
-        else f"Selected task: {task_id} ({status})"
-    )
+    header = (
+        f"Selected task: none\nStatus: {status}"
+        if task_id is None
+        else f"Selected task: {task_id} ({status})"
+    )
     return f"<workflow-state>\n{header}\n{body}\n</workflow-state>"
 
 
@@ -323,15 +323,15 @@ def main() -> int:
     templates = load_breadcrumbs(root)
     platform = _detect_platform(data)
     config = _read_trellis_config(root)
-    task = get_selected_task(root, data)
+    task = get_selected_task(root, data)
     if task is None:
-        # No selected task — still emit the framework-entry breadcrumb.
-        no_task_key = resolve_breadcrumb_key("no_task", platform, config)
+        # No selected task — still emit the framework-entry breadcrumb.
+        no_task_key = resolve_breadcrumb_key("no_task", platform, config)
         breadcrumb = build_breadcrumb(
             None, "no_task", templates, breadcrumb_key=no_task_key
         )
     else:
-        task_id, status, source = task
+        task_id, status, source = task
         status_key = resolve_breadcrumb_key(status, platform, config)
         source_for_breadcrumb = None if platform == "codex" else source
         breadcrumb = build_breadcrumb(

package/dist/templates/shared-hooks/research-end-retrieval-pack.py

@@ -33,7 +33,7 @@ if sys.platform.startswith("win"):
 
 DIR_WORKFLOW = ".trellis"
 OUTPUT_BASENAME = "retrieval-pack-latest.json"
-MARKER = "<!-- trellis-research-end-pack -->"
+MARKER = "<!-- cstl-research-end-pack -->"
 
 
 def _repo_root(cwd: str) -> Path | None:

package/dist/templates/shared-hooks/session-start.py

@@ -1,4 +1,4 @@
-#!/usr/bin/env python3
+#!/usr/bin/env python3
 # -*- coding: utf-8 -*-
 """
 Session Start Hook - Inject structured context
@@ -365,7 +365,7 @@ def _get_task_status(trellis_dir: Path, input_data: dict) -> str:
         return (
             f"Status: COMPLETED\nTask: {task_title}\n"
             f"Present: {present_line}\n"
-            "Next-Action: Run `/trellis:finish-work`. If the working tree is dirty, return to Phase 3.4 first."
+            "Next-Action: Run `/cstl:finish-work`. If the working tree is dirty, return to Phase 3.4 first."
         )
 
     has_prd = (task_dir / "prd.md").is_file()
@@ -382,7 +382,7 @@ def _get_task_status(trellis_dir: Path, input_data: dict) -> str:
         return (
             f"Status: PLANNING\nTask: {task_title}\n"
             f"Present: {present_line}\n"
-            "Next-Action: Load `trellis-brainstorm` and write `prd.md`. Stay in planning."
+            "Next-Action: Load `cstl-brainstorm` and write `prd.md`. Stay in planning."
         )
 
     if task_status == "planning":

package/dist/templates/trellis/config/execution-strategy-rules.json

@@ -20,7 +20,7 @@
     {
       "execution_mode": "worker",
       "isolation": "git-worktree",
-      "message": "worker usually runs in main-worktree; git-worktree is rare for trellis-implement"
+      "message": "worker usually runs in main-worktree; git-worktree is rare for cstl-implement"
     },
     {
       "execution_mode": "inline",

package/dist/templates/trellis/config.yaml

@@ -55,7 +55,7 @@ max_journal_lines: 2000
 #-------------------------------------------------------------------------------
 
 # Declare packages for monorepo projects.
-# Trellis auto-detects workspaces during `trellis init`, but you can also
+# Trellis auto-detects workspaces during `cstl init`, but you can also
 # configure them manually here.
 #
 # packages:
@@ -101,7 +101,7 @@ max_journal_lines: 2000
 #-------------------------------------------------------------------------------
 # Channel worker OOM guard
 #-------------------------------------------------------------------------------
-# Default safeguards for `trellis channel spawn` workers. The guard runs
+# Default safeguards for `cstl channel spawn` workers. The guard runs
 # at spawn time (cleans expired idle workers, then enforces the live-worker
 # budget) and inside each supervisor (self-terminates a worker that stays
 # continuously idle past `idle_timeout`).
@@ -125,8 +125,8 @@ channel:
 # main Codex agent edit code directly because Codex sub-agents run with
 # `fork_turns="none"` isolation and can't inherit the parent session's
 # task context. Set to "sub-agent" to opt into the legacy dispatch model
-# (main agent spawns trellis-implement / trellis-check / trellis-research
+# (main agent spawns cstl-implement / cstl-check / cstl-research
 # sub-agents).
 #
 # codex:
-#   dispatch_mode: inline  # or "sub-agent" to dispatch trellis-* sub-agents
+#   dispatch_mode: inline  # or "sub-agent" to dispatch cstl-* sub-agents

package/dist/templates/trellis/local/README.md

@@ -1,8 +1,10 @@
 # Cursor++ Local Bundle
 
-> **Native Cursor users:** safe to ignore or delete this directory unless using **Cursor++ BYOK**.
+> **Native Cursor users:** safe to ignore or delete this directory unless using **Cursor++ BYOK** on **this repo**.
 
-**Who needs this:** Cursor IDE with [Cursor++](https://ccursor.cometix.dev) (BYOK proxy). Method 2.5 patches the `WPeLc8` resolver so Trellis Task subagents get per-role BYOK slugs instead of inheriting the parent model.
+**Coexistence:** Native and BYOK can run on the same machine across different projects. This bundle is **per-repo** (`cstl init --cursor --cursor2plus`); sibling repos may use `cstl init --cursor` only. Full guide: cursor-trellis `docs/cursor.md` — section *Native and BYOK coexistence (not either/or)*.
+
+**Who needs this:** Cursor IDE with [Cursor++](https://ccursor.cometix.dev) (BYOK proxy). Method 2.5 patches the Cursor++ task model resolver (historically located via `WPeLc8` in verified builds) so Trellis Task subagents get per-role BYOK slugs instead of inheriting the parent model. **After any Cursor or Cursor++ upgrade**, run `python patch_wpelc8.py --check-compat` before re-applying.
 
 ## Bundle files
 
@@ -28,10 +30,10 @@ User-wide model doc: `~/.ccursor/trellis-task-models.json5` (see `../trellis-tas
 2. **Operator approval** — review map and compat output; decide to proceed
 3. **Apply** — `python patch_wpelc8.py --apply --approve` (writes `extension.js`)
 4. **Reload Cursor** — Developer: Reload Window (once per apply/revert)
-5. **A ≠ B probe** — parent chat model must differ from a patched subagent map target; dispatch a `trellis-*` Task and confirm the subagent self-reports the **map target** slug/model, not the parent
+5. **A ≠ B probe** — parent chat model must differ from a patched subagent map target; dispatch a `cstl-*` Task and confirm the subagent self-reports the **map target** slug/model, not the parent
 6. **Revert if needed** — `python patch_wpelc8.py --revert` then Reload Window
 
-Agent-led setup: run skill **`trellis-cursor2plus-setup`** in Cursor Agent (lists models, writes JSON5, runs patch steps).
+Agent-led setup: run skill **`cstl-cursor2plus-setup`** in Cursor Agent (lists models, writes JSON5, runs patch steps).
 
 ## Revert recipe
 
@@ -47,10 +49,10 @@ Then **Developer: Reload Window**. Subagents return to **inherit parent** BYOK m
 Goal: prove the patch routes **subagent** model independently of the **parent** session model.
 
 1. Set parent chat to model **A** (any BYOK model in Cursor++).
-2. Ensure `trellis-task-models.json5` maps e.g. `trellis-implement` to model **B** (`primary`/`fallback` resolving to a different slug than A).
-3. `python patch_wpelc8.py --print-map` — note the slug for `trellis-implement` (= map target **B**).
+2. Ensure `trellis-task-models.json5` maps e.g. `cstl-implement` to model **B** (`primary`/`fallback` resolving to a different slug than A).
+3. `python patch_wpelc8.py --print-map` — note the slug for `cstl-implement` (= map target **B**).
 4. `python patch_wpelc8.py --apply --approve` → Reload Window.
-5. Dispatch a `trellis-implement` Task from a parent on model **A**.
+5. Dispatch a `cstl-implement` Task from a parent on model **A**.
 6. **Pass:** subagent self-report / trace shows **B** (map target), not **A**.
 7. **Fail:** subagent shows **A** → patch stale or not applied; run `--check-compat`, re-patch, or `--revert`.
 
@@ -73,9 +75,9 @@ python smoke.py                             # env + patch_status smoke
 
 | JSON key (under `models`) | Meaning |
 |---------------------------|---------|
-| `trellis-research` | Trellis research Task subagent |
-| `trellis-implement` | Trellis implement (+ Parent/Child default Task) |
-| `trellis-check` | Trellis check Task subagent |
+| `cstl-research` | Trellis research Task subagent |
+| `cstl-implement` | Trellis implement (+ Parent/Child default Task) |
+| `cstl-check` | Trellis check Task subagent |
 | `generalPurpose` / `shell` / `best-of-n-runner` | Cursor built-in Task types |
 
 Each role: `{ "primary": "<name>", "fallback": "<name>" }`. Names: **apiModel** or **displayName** from Cursor++ (`--list-models`). **Explore** is not in this file — use the Cursor++ panel.

package/dist/templates/trellis/local/patch_wpelc8.py

@@ -166,7 +166,7 @@ def try_resolve_extension_js(bundle_dir: Path, config: dict) -> Path | None:
 def build_inject_block(model_map: dict[str, str]) -> str:
     if not model_map:
         raise SystemExit(
-            "model map empty — use skill trellis-cursor2plus-setup or write "
+            "model map empty — use skill cstl-cursor2plus-setup or write "
             "~/.ccursor/trellis-task-models.json5 (see .trellis/local/trellis-task-models.json5.example)"
         )
     pairs = ",".join(f'"{k}":"{v}"' for k, v in sorted(model_map.items()))
@@ -240,7 +240,7 @@ def cmd_explain() -> None:
         print(f"  {key}")
         print(f"    → {SUBAGENT_TYPE_HELP.get(key, '')}")
     print()
-    print("Skill: trellis-cursor2plus-setup (init 后与大模型交互完成配置)")
+    print("Skill: cstl-cursor2plus-setup (init 后与大模型交互完成配置)")
 
 
 def cmd_list_models(providers_path: Path) -> None:
@@ -272,7 +272,7 @@ def cmd_bootstrap(bundle_dir: Path) -> None:
             config["extensionJs"] = path
     cfg_out.write_text(json.dumps(config, indent=2) + "\n", encoding="utf-8")
     print("wrote", cfg_out)
-    print("Next: in Cursor Agent, run skill trellis-cursor2plus-setup", file=sys.stderr)
+    print("Next: in Cursor Agent, run skill cstl-cursor2plus-setup", file=sys.stderr)
 
 
 def cmd_check_compat(bundle_dir: Path, config: dict) -> int:

package/dist/templates/trellis/local/subagent-models.json.example

@@ -1,4 +1,4 @@
 {
   "_comment": "Optional repo overrides (~/.ccursor/trellis-task-models.json is user-wide). Same value rules: apiModel, displayName, or model-xxxxx slug. Project wins on same key.",
-  "trellis-implement": "qwen3.6-plus"
+  "cstl-implement": "qwen3.6-plus"
 }

package/dist/templates/trellis/local/trellis-task-models.json.example

@@ -1,9 +1,9 @@
 {
   "_comment": "Copy to ~/.ccursor/trellis-task-models.json (Windows: %USERPROFILE%\\.ccursor\\). Values: use apiModel or displayName from Cursor++ — patch_wpelc8.py resolves to slug at patch time. Optional keys: omit types you do not use.",
   "_when_providers_change": "Run: python patch_wpelc8.py --list-models then fix this file and patch again.",
-  "trellis-research": "glm-5.1",
-  "trellis-implement": "qwen3.6-plus",
-  "trellis-check": "qwen3.7-max",
+  "cstl-research": "glm-5.1",
+  "cstl-implement": "qwen3.6-plus",
+  "cstl-check": "qwen3.7-max",
   "generalPurpose": "gpt-5.5",
   "shell": "grok-composer-2.5-fast",
   "best-of-n-runner": "kimi-k2.7-code"

package/dist/templates/trellis/local/trellis-task-models.json5.example

@@ -10,20 +10,20 @@
 
 {
   "models": {
-    // Trellis 调研（Task trellis-research）
-    "trellis-research": {
+    // Trellis 调研（Task cstl-research）
+    "cstl-research": {
       "primary": "glm-5.1",
       "fallback": "qwen3.7-max"
     },
 
     // Trellis 实现；Parent generate-child-prompt --mode subagent 默认派发此类型
-    "trellis-implement": {
+    "cstl-implement": {
       "primary": "qwen3.6-plus",
       "fallback": "kimi-k2.7-code"
     },
 
-    // Trellis 验收（Task trellis-check）
-    "trellis-check": {
+    // Trellis 验收（Task cstl-check）
+    "cstl-check": {
       "primary": "qwen3.7-max",
       "fallback": "glm-5.1"
     },

package/dist/templates/trellis/local/trellis_task_models_config.py

@@ -6,18 +6,18 @@ from pathlib import Path
 from typing import Any
 
 KNOWN_SUBAGENT_TYPES = (
-    "trellis-research",
-    "trellis-implement",
-    "trellis-check",
+    "cstl-research",
+    "cstl-implement",
+    "cstl-check",
     "generalPurpose",
     "shell",
     "best-of-n-runner",
 )
 
 SUBAGENT_TYPE_HELP: dict[str, str] = {
-    "trellis-research": "Trellis 调研 Task（trellis-research）",
-    "trellis-implement": "Trellis 实现 Task；Parent/Child 默认派发类型",
-    "trellis-check": "Trellis 验收 Task（trellis-check）",
+    "cstl-research": "Trellis 调研 Task（cstl-research）",
+    "cstl-implement": "Trellis 实现 Task；Parent/Child 默认派发类型",
+    "cstl-check": "Trellis 验收 Task（cstl-check）",
     "generalPurpose": "Cursor 内置 Task：generalPurpose",
     "shell": "Cursor 内置 Task：shell",
     "best-of-n-runner": "Cursor 内置 Task：best-of-n-runner",

package/dist/templates/trellis/scripts/common/cli_adapter.py

@@ -46,15 +46,15 @@ class CLIAdapter:
     def get_commands_path(self, project_root: Path, *parts: str) -> Path:
         if not parts:
             return self.get_config_dir(project_root) / "commands"
-        if len(parts) >= 2 and parts[0] == "trellis":
+        if len(parts) >= 2 and parts[0] == "cstl":
             filename = parts[-1]
             return (
-                self.get_config_dir(project_root) / "commands" / f"trellis-{filename}"
+                self.get_config_dir(project_root) / "commands" / f"cstl-{filename}"
             )
         return self.get_config_dir(project_root) / "commands" / Path(*parts)
 
     def get_trellis_command_path(self, name: str) -> str:
-        return f".cursor/commands/trellis-{name}.md"
+        return f".cursor/commands/cstl-{name}.md"
 
     def get_non_interactive_env(self) -> dict[str, str]:
         return {}

package/dist/templates/trellis/scripts/common/execution_strategy.py

@@ -195,7 +195,7 @@ def suggest_execution_strategy(
     elif profile == "full":
         mode = "worker"
         iso = "main-worktree"
-        reasons.append("Full task touching code → worker (trellis-implement / check sub-agents)")
+        reasons.append("Full task touching code → worker (cstl-implement / check sub-agents)")
     else:
         mode = "inline"
         iso = "main-worktree"