@blxzer/cursor-trellis 0.3.0 → 0.3.3

package/dist/templates/trellis/tasks/templates/release-readiness/implement.md

@@ -15,13 +15,13 @@ quality_gates:
 
 ## Steps
 
-1. Read product release spec (e.g. `.trellis/spec/**/release*.md`) and recent integrated task handoffs.
+1. Read product release spec (e.g. `.cstl/spec/**/release*.md`) and recent integrated task handoffs.
 2. Record current versions from `package.json` (and lockstep packages).
 3. Draft version recommendation and changelog.
 4. Run non-mutating guards and validation (manifest, build, focused tests, typecheck, pack dry-run).
 5. Fill `verify.md` using the evidence contract in `design.md`.
 6. Write `handoff.md` with **Ready to publish**, **Not published**, and **Blockers** tables.
-7. Run `python ./.trellis/scripts/task.py validate <this-task>` and `archive --check` when finishing.
+7. Run `python ./.cstl/scripts/task.py validate <this-task>` and `archive --check` when finishing.
 
 ## Validation
 

package/dist/templates/trellis/workflow.md

@@ -19,96 +19,96 @@
 On first use, initialize your identity:
 
 ```bash
-python3 ./.trellis/scripts/init_developer.py <your-name>
+python3 ./.cstl/scripts/init_developer.py <your-name>
 ```
 
-Creates `.trellis/.developer` (gitignored) + `.trellis/workspace/<your-name>/`.
+Creates `.cstl/.developer` (gitignored) + `.cstl/workspace/<your-name>/`.
 
 ### Spec System
 
-`.trellis/spec/` holds coding guidelines organized by package and layer.
+`.cstl/spec/` holds coding guidelines organized by package and layer.
 
-- `.trellis/spec/<package>/<layer>/index.md` — entry point with **Pre-Development Checklist** + **Quality Check**. Actual guidelines live in the `.md` files it points to.
-- `.trellis/spec/guides/index.md` — cross-package thinking guides.
+- `.cstl/spec/<package>/<layer>/index.md` — entry point with **Pre-Development Checklist** + **Quality Check**. Actual guidelines live in the `.md` files it points to.
+- `.cstl/spec/guides/index.md` — cross-package thinking guides.
 
 ```bash
-python3 ./.trellis/scripts/get_context.py --mode packages   # list packages / layers
+python3 ./.cstl/scripts/get_context.py --mode packages   # list packages / layers
 ```
 
 **When to update spec**: new pattern/convention found · bug-fix prevention to codify · new technical decision.
 
 ### Task System
 
-Every task has its own directory under `.trellis/tasks/{MM-DD-name}/` holding `task.json`, `prd.md`, optional `design.md`, optional `implement.md`, optional `research/`, and context manifests (`implement.jsonl`, `check.jsonl`) for Cursor (sub-agent dispatch).
+Every task has its own directory under `.cstl/tasks/{MM-DD-name}/` holding `task.json`, `prd.md`, optional `design.md`, optional `implement.md`, optional `research/`, and context manifests (`implement.jsonl`, `check.jsonl`) for Cursor (sub-agent dispatch).
 
 ```bash
 # Task lifecycle
-python3 ./.trellis/scripts/task.py create "<title>" [--slug <name>] [--parent <dir>]
-python3 ./.trellis/scripts/task.py dashboard             # show Task Dashboard without mutating state
-python3 ./.trellis/scripts/task.py select <name>         # select task for this live session
-python3 ./.trellis/scripts/task.py selected --source     # show selected task and source
-python3 ./.trellis/scripts/task.py start-execution <name> --check
-python3 ./.trellis/scripts/task.py start-execution <name> --approved
-python3 ./.trellis/scripts/task.py exit                  # clear selected task without changing status
-python3 ./.trellis/scripts/task.py archive <name>        # move to archive/{year-month}/
-python3 ./.trellis/scripts/task.py list [--mine] [--status <s>]
-python3 ./.trellis/scripts/task.py list-archive
-python3 ./.trellis/scripts/task.py add-subtask <parent> <child>
-python3 ./.trellis/scripts/task.py set-child-state <parent> <child> review --evidence verify.md
-python3 ./.trellis/scripts/task.py prepare-child-worktree <parent> <child> --branch <child-branch>
-python3 ./.trellis/scripts/task.py integrate-child <parent> <child> accepted --evidence handoff.md --ref <child-ref>
+python3 ./.cstl/scripts/task.py create "<title>" [--slug <name>] [--parent <dir>]
+python3 ./.cstl/scripts/task.py dashboard             # show Task Dashboard without mutating state
+python3 ./.cstl/scripts/task.py select <name>         # select task for this live session
+python3 ./.cstl/scripts/task.py selected --source     # show selected task and source
+python3 ./.cstl/scripts/task.py start-execution <name> --check
+python3 ./.cstl/scripts/task.py start-execution <name> --approved
+python3 ./.cstl/scripts/task.py exit                  # clear selected task without changing status
+python3 ./.cstl/scripts/task.py archive <name>        # move to archive/{year-month}/
+python3 ./.cstl/scripts/task.py list [--mine] [--status <s>]
+python3 ./.cstl/scripts/task.py list-archive
+python3 ./.cstl/scripts/task.py add-subtask <parent> <child>
+python3 ./.cstl/scripts/task.py set-child-state <parent> <child> review --evidence verify.md
+python3 ./.cstl/scripts/task.py prepare-child-worktree <parent> <child> --branch <child-branch>
+python3 ./.cstl/scripts/task.py integrate-child <parent> <child> accepted --evidence handoff.md --ref <child-ref>
 
 # Code-spec context (injected into implement/check agents via JSONL).
 # `implement.jsonl` / `check.jsonl` are seeded on `task create` for sub-agent-capable
 # platforms; the AI curates real spec + research entries during planning when needed.
-python3 ./.trellis/scripts/task.py add-context <name> <action> <file> <reason>
-python3 ./.trellis/scripts/task.py list-context <name> [action]
-python3 ./.trellis/scripts/task.py validate <name>
+python3 ./.cstl/scripts/task.py add-context <name> <action> <file> <reason>
+python3 ./.cstl/scripts/task.py list-context <name> [action]
+python3 ./.cstl/scripts/task.py validate <name>
 
 # Task metadata
-python3 ./.trellis/scripts/task.py set-branch <name> <branch>
-python3 ./.trellis/scripts/task.py set-base-branch <name> <branch>    # PR target
-python3 ./.trellis/scripts/task.py set-scope <name> <scope>
+python3 ./.cstl/scripts/task.py set-branch <name> <branch>
+python3 ./.cstl/scripts/task.py set-base-branch <name> <branch>    # PR target
+python3 ./.cstl/scripts/task.py set-scope <name> <scope>
 
 # Hierarchy (parent/child)
-python3 ./.trellis/scripts/task.py add-subtask <parent> <child>
-python3 ./.trellis/scripts/task.py remove-subtask <parent> <child>
+python3 ./.cstl/scripts/task.py add-subtask <parent> <child>
+python3 ./.cstl/scripts/task.py remove-subtask <parent> <child>
 
 # PR creation
-python3 ./.trellis/scripts/task.py create-pr [name] [--dry-run]
+python3 ./.cstl/scripts/task.py create-pr [name] [--dry-run]
 ```
 
-> Run `python3 ./.trellis/scripts/task.py --help` to see the authoritative, up-to-date list.
+> Run `python3 ./.cstl/scripts/task.py --help` to see the authoritative, up-to-date list.
 
 **Selected-task mechanism**: entering a Trellis project activates framework context, but every new live session starts with `Selected task: none`. `task.py create` creates artifacts only. `task.py select <task>` writes a per-session `selected_task` pointer without changing `task.json.status`. `task.py selected --source` reports that pointer. `task.py exit` clears it without changing status. `task.py start-execution <task> --check` verifies execution readiness without mutation; `task.py start-execution <task> --approved` is the explicit execution boundary and may flip `planning` to `in_progress`. `task.py archive <task>` writes `status=completed`, moves the directory to `archive/`, and deletes runtime session files that still point at the archived task.
 
 ### Workspace System
 
-Records every AI session for cross-session tracking under `.trellis/workspace/<developer>/`.
+Records every AI session for cross-session tracking under `.cstl/workspace/<developer>/`.
 
 - `journal-N.md` — session log. **Max 2000 lines per file**; a new `journal-(N+1).md` is auto-created when exceeded.
 - `index.md` — personal index (total sessions, last active).
 
 ```bash
-python3 ./.trellis/scripts/add_session.py --title "Title" --commit "hash" --summary "Summary"
+python3 ./.cstl/scripts/add_session.py --title "Title" --commit "hash" --summary "Summary"
 ```
 
 ### Context Script
 
 ```bash
-python3 ./.trellis/scripts/get_context.py                            # full session runtime
-python3 ./.trellis/scripts/get_context.py --mode packages            # available packages + spec layers
-python3 ./.trellis/scripts/get_context.py --mode phase --step <X.Y>  # detailed guide for a workflow step
-python3 ./.trellis/scripts/get_context.py --mode retrieval-pack --json --input <evidence.json>  # score collected evidence (not default --json)
+python3 ./.cstl/scripts/get_context.py                            # full session runtime
+python3 ./.cstl/scripts/get_context.py --mode packages            # available packages + spec layers
+python3 ./.cstl/scripts/get_context.py --mode phase --step <X.Y>  # detailed guide for a workflow step
+python3 ./.cstl/scripts/get_context.py --mode retrieval-pack --json --input <evidence.json>  # score collected evidence (not default --json)
 ```
 
 **Evidence scoring:** default `--json` returns `retrievalGuide` only. After collecting artifact search, session memory, smart-search manifests under `{TASK}/research/smart-search/`, or codebase candidates, run **`--mode retrieval-pack`** with `--input` or stdin JSON. See `research/evidence-scoring-integration.md` in the active task or archived `06-15-child-phase2-evidence-scoring`.
 
 **Research-end hook (Cursor `stop`):** when the selected task has `{TASK}/research/*.md` or `research/smart-search/`, `.cursor/hooks/research-end-retrieval-pack.py` may write `{TASK}/research/retrieval-pack-latest.json` via `get_context --mode retrieval-pack`. Default session JSON is unchanged; use the file when closing research or before Phase 3.1.
 
-**Retrieval daily guide:** `.trellis/spec/guides/retrieval-daily-guide.md` — when to use rg, codegraph, fast-context-mcp, smart-search-cli (and Cursor web fallback), artifact/session memory, codebase router (suggest-only), and explicit retrieval-pack scoring.
+**Retrieval daily guide:** `.cstl/spec/guides/retrieval-daily-guide.md` — when to use rg, codegraph, fast-context-mcp, smart-search-cli (and Cursor web fallback), artifact/session memory, codebase router (suggest-only), and explicit retrieval-pack scoring.
 
-**Cursor subagent dispatch:** `.trellis/spec/guides/cursor-subagent-policy.md` — `cstl-research` / `cstl-implement` / `cstl-check`; Parent child default **Task** `cstl-implement` from Parent session (`generate-child-prompt --mode subagent`). **Cursor++ BYOK:** per-type models via `.trellis/local/cursor2plus/` + user/project JSON maps (not committed slugs). **Native Cursor API:** frontmatter `model:` on agents still works. PRD Grill stays in `cstl-brainstorm`, not a subagent. **Cursor++:** compatible v0.0.11+ (SubAgent readonly bug fixed).
+**Cursor subagent dispatch:** `.cstl/spec/guides/cursor-subagent-policy.md` — `cstl-research` / `cstl-implement` / `cstl-check`; Parent child default **Task** `cstl-implement` from Parent session (`generate-child-prompt --mode subagent`). **Cursor++ BYOK:** per-type models via `.cstl/local/cursor2plus/` + user/project JSON maps (not committed slugs). **Native Cursor API:** frontmatter `model:` on agents still works. PRD Grill stays in `cstl-brainstorm`, not a subagent. **Cursor++:** compatible v0.0.11+ (SubAgent readonly bug fixed).
 
 ---
 
@@ -151,7 +151,7 @@ python3 ./.trellis/scripts/get_context.py --mode retrieval-pack --json --input <
     - Run `cstl update` after editing to push the new bodies to
       downstream user projects (block-level managed replacement)
     - Full runtime contract:
-      .trellis/spec/cli/backend/workflow-state-contract.md
+      .cstl/spec/cli/backend/workflow-state-contract.md
 -->
 
 ## Phase Index
@@ -223,7 +223,7 @@ Before executing an upgrade that creates artifacts, changes task mode, adds gate
 | No selected task + need dashboard | `cstl-start` |
 | Selected task + resume step | `cstl-continue` |
 | Planning / PRD | `cstl-brainstorm` |
-| Parent with parallel children | `generate-child-prompt --mode subagent`; writable Agent; see `.trellis/spec/guides/cursor-subagent-policy.md` |
+| Parent with parallel children | `generate-child-prompt --mode subagent`; writable Agent; see `.cstl/spec/guides/cursor-subagent-policy.md` |
 
 Details: archived `06-15-child-phase3-task-ladder` → `research/task-ladder-iteration.md`.
 
@@ -254,11 +254,11 @@ Integration is Parent/Child-only. Ordinary Lite and Full Tasks skip Integration
 Parent sessions can productize child dispatch and review without a new agent runtime:
 
 ```bash
-python3 ./.trellis/scripts/task.py parent-status <parent-task>
-python3 ./.trellis/scripts/task.py generate-child-prompt <parent-task> <child-task> --mode inline
-python3 ./.trellis/scripts/task.py review-child <parent-task> <child-task> --check --decision accept --ref <child-ref>
-python3 ./.trellis/scripts/task.py review-child <parent-task> <child-task> --decision accept --ref <child-ref>
-python3 ./.trellis/scripts/task.py review-child <parent-task> <child-task> --decision integrate-through --ref <child-ref>
+python3 ./.cstl/scripts/task.py parent-status <parent-task>
+python3 ./.cstl/scripts/task.py generate-child-prompt <parent-task> <child-task> --mode inline
+python3 ./.cstl/scripts/task.py review-child <parent-task> <child-task> --check --decision accept --ref <child-ref>
+python3 ./.cstl/scripts/task.py review-child <parent-task> <child-task> --decision accept --ref <child-ref>
+python3 ./.cstl/scripts/task.py review-child <parent-task> <child-task> --decision integrate-through --ref <child-ref>
 ```
 
 - `generate-child-prompt` reads parent `task-map.md` for `depends_on` and `touches`, child artifacts, and optional parent `child-prompts.md`. Use `--mode subagent` only as a delivery hint when the platform can spawn subagents; inline mode remains the portable default.
@@ -380,8 +380,8 @@ Implementation defects inside the already approved contract route back to Execut
 At each step, run this to fetch detailed guidance:
 
 ```bash
-python3 ./.trellis/scripts/get_context.py --mode phase --step <step>
-# e.g. python3 ./.trellis/scripts/get_context.py --mode phase --step 1.1
+python3 ./.cstl/scripts/get_context.py --mode phase --step <step>
+# e.g. python3 ./.cstl/scripts/get_context.py --mode phase --step 1.1
 ```
 
 ---
@@ -395,7 +395,7 @@ Goal: classify the request, get task-creation consent when a task is needed, and
 Create the task directory only after task-creation consent. The command sets status to `planning`, writes `task.json`, and creates a default `prd.md`. It does not select the task, start execution, or approve execution:
 
 ```bash
-python3 ./.trellis/scripts/task.py create "<task title>" --slug <name>
+python3 ./.cstl/scripts/task.py create "<task title>" --slug <name>
 ```
 
 `--slug` is the human-readable name only. Do **not** include the `MM-DD-` date prefix; `task.py create` adds that prefix automatically.
@@ -405,14 +405,14 @@ For task trees, create the parent task first and then create each child with `--
 After creation, select the task only when the user has chosen it for this live session:
 
 ```bash
-python3 ./.trellis/scripts/task.py select <task-dir>
+python3 ./.cstl/scripts/task.py select <task-dir>
 ```
 
 After selection, the per-turn breadcrumb switches to `[workflow-state:planning]`, telling the AI to stay in planning.
 
 Run only `create` and, when appropriate, `select` here. Do not run `start-execution --approved` until step 1.4 passes its non-mutating `--check` and the user gives explicit execution approval.
 
-Skip when the user has already explicitly selected an appropriate task with `python3 ./.trellis/scripts/task.py select <task>`.
+Skip when the user has already explicitly selected an appropriate task with `python3 ./.cstl/scripts/task.py select <task>`.
 
 #### 1.1 Requirement exploration `[required · repeatable]`
 
@@ -450,7 +450,7 @@ Spawn the research sub-agent:
 
 
 **Retrieval during research**:
-- Use `python3 ./.trellis/scripts/search_artifacts.py --query "<topic>" --json` to find durable Trellis specs, prior tasks, research, verification notes, and workspace journals before re-discovering framework context.
+- Use `python3 ./.cstl/scripts/search_artifacts.py --query "<topic>" --json` to find durable Trellis specs, prior tasks, research, verification notes, and workspace journals before re-discovering framework context.
 - Use `codebase-retrieval` evidence levels for source-code questions: adapter output is candidate evidence until current source, Git, or validation confirms it.
 - Persist useful exploratory retrieval chains, adapter availability, and competing hypotheses under `{TASK_DIR}/research/`.
 
@@ -486,7 +486,7 @@ Curate `implement.jsonl` and `check.jsonl` so the Phase 2 sub-agents get the rig
 **Format**: one JSON object per line — `{"file": "<path>", "reason": "<why>"}`. Paths are repo-root relative.
 
 **What to put in**:
-- **Spec files** — `.trellis/spec/<package>/<layer>/index.md` and any specific guideline files (`error-handling.md`, `conventions.md`, etc.) relevant to this task
+- **Spec files** — `.cstl/spec/<package>/<layer>/index.md` and any specific guideline files (`error-handling.md`, `conventions.md`, etc.) relevant to this task
 - **Research files** — `{TASK_DIR}/research/*.md` that the sub-agent will need to consult
 
 **What NOT to put in**:
@@ -502,7 +502,7 @@ These manifests do not replace `implement.md`. `implement.md` is the human-reada
 **How to discover relevant specs**:
 
 ```bash
-python3 ./.trellis/scripts/get_context.py --mode packages
+python3 ./.cstl/scripts/get_context.py --mode packages
 ```
 
 Lists every package + its spec layers with paths. Pick the entries that match this task's domain.
@@ -510,7 +510,7 @@ Lists every package + its spec layers with paths. Pick the entries that match th
 Use artifact search when prior task/research evidence is likely relevant:
 
 ```bash
-python3 ./.trellis/scripts/search_artifacts.py --query "<topic>" --json
+python3 ./.cstl/scripts/search_artifacts.py --query "<topic>" --json
 ```
 
 Add any reusable `{TASK_DIR}/research/*.md` files you discovered when sub-agents need them.
@@ -520,8 +520,8 @@ Add any reusable `{TASK_DIR}/research/*.md` files you discovered when sub-agents
 Either edit the jsonl file directly in your editor, or use:
 
 ```bash
-python3 ./.trellis/scripts/task.py add-context "$TASK_DIR" implement "<path>" "<reason>"
-python3 ./.trellis/scripts/task.py add-context "$TASK_DIR" check "<path>" "<reason>"
+python3 ./.cstl/scripts/task.py add-context "$TASK_DIR" implement "<path>" "<reason>"
+python3 ./.cstl/scripts/task.py add-context "$TASK_DIR" check "<path>" "<reason>"
 ```
 
 Delete the seed `_example` line once real entries exist (optional — it's skipped automatically by consumers).
@@ -535,7 +535,7 @@ Skip when: `implement.jsonl` and `check.jsonl` have agent-curated entries (the s
 After artifact review, run the non-mutating execution preflight:
 
 ```bash
-python3 ./.trellis/scripts/task.py start-execution <task-dir> --check
+python3 ./.cstl/scripts/task.py start-execution <task-dir> --check
 ```
 
 For lightweight tasks, `prd.md` can be enough. For complex tasks, `prd.md`, `design.md`, and `implement.md` must exist and be reviewed before execution approval. On Cursor (sub-agent dispatch), curate jsonl manifests when extra spec or research context is needed; seed-only manifests are tolerated by consumers.
@@ -545,7 +545,7 @@ If `--check` passes, report that artifact gates are ready and include the task n
 Only after the user approves execution in that context, run:
 
 ```bash
-python3 ./.trellis/scripts/task.py start-execution <task-dir> --approved
+python3 ./.cstl/scripts/task.py start-execution <task-dir> --approved
 ```
 
 After this command succeeds, the breadcrumb switches to `[workflow-state:in_progress]`, and the rest of Phase 2 / 3 follows.
@@ -581,7 +581,7 @@ Implementation defects inside the approved contract remain Execution work: fix t
 #### 2.1 Implement `[required · repeatable]`
 
 Use retrieval layers before and during implementation when context is incomplete:
-- `python3 ./.trellis/scripts/search_artifacts.py --query "<topic>" --json` for durable Trellis specs, prior task artifacts, research, verification notes, and journals.
+- `python3 ./.cstl/scripts/search_artifacts.py --query "<topic>" --json` for durable Trellis specs, prior task artifacts, research, verification notes, and journals.
 - `codebase-retrieval` evidence levels for source claims: candidate -> corroborated candidate -> verified claim; unresolved or unavailable adapters must be reported instead of treated as proof.
 - Record exploratory chains in `{TASK_DIR}/research/` and final source/Git/test proof in `verify.md`.
 
@@ -629,7 +629,7 @@ Goal: ensure code quality, capture lessons, record the work.
 
 For **complex** tasks, optionally before 3.1 when you have collected retrieval evidence JSON (e.g. after Phase 1.2 smart-search manifests in `{TASK}/research/smart-search/`):
 
-`python3 ./.trellis/scripts/get_context.py --mode retrieval-pack --json --input <path-to-evidence.json>`
+`python3 ./.cstl/scripts/get_context.py --mode retrieval-pack --json --input <path-to-evidence.json>`
 
 Use `contextPack.selected` / `scoredEvidence` to order citations in `verify.md`. Do **not** expect scoring in default `get_context --json`.
 
@@ -663,7 +663,7 @@ Review whether this task produced durable learning worth recording:
 - requirement drift, architecture decisions, reusable conventions, or toolchain pitfalls;
 - new project-local patterns that should affect future work.
 
-If durable learning exists, load `cstl-update-spec` and update `.trellis/spec/` or write a focused `retrospective.md`, then link that evidence from `verify.md`.
+If durable learning exists, load `cstl-update-spec` and update `.cstl/spec/` or write a focused `retrospective.md`, then link that evidence from `verify.md`.
 
 If no durable learning exists, write an explicit `No durable learning` decision in `verify.md`. Do not run a spec update only to satisfy ceremony.
 
@@ -801,5 +801,5 @@ Supported events: `after_create / after_start / after_archive`. Use `after_archi
 
 For the workflow state machine's runtime contract, the locations of all status writers, pseudo-statuses (`no_task` / `stale_<source_type>`), the hook reachability matrix, and other deep details, see:
 
-- `.trellis/spec/cli/backend/workflow-state-contract.md` — runtime contract + writer table + test invariants
-- `.trellis/scripts/inject-workflow-state.py` — actual parser (reads workflow.md only, no embedded text)
+- `.cstl/spec/cli/backend/workflow-state-contract.md` — runtime contract + writer table + test invariants
+- `.cstl/scripts/inject-workflow-state.py` — actual parser (reads workflow.md only, no embedded text)

package/dist/types/migration.d.ts

@@ -35,7 +35,7 @@ export interface MigrationItem {
  * substring (live or commented), the section is treated as already present.
  */
 export interface ConfigSectionAdded {
-    /** Target file relative to project root (e.g. `.trellis/config.yaml`). */
+    /** Target file relative to project root (e.g. `.cstl/config.yaml`). */
     file: string;
     /**
      * Substring whose presence in the user file means this section already

package/dist/types/migration.d.ts.map

@@ -1 +1 @@
-{"version":3,"file":"migration.d.ts","sourceRoot":"","sources":["../../src/types/migration.ts"],"names":[],"mappings":"AAAA;;;;;GAKG;AAEH;;GAEG;AACH,MAAM,WAAW,aAAa;IAC5B,+BAA+B;IAC/B,IAAI,EAAE,QAAQ,GAAG,YAAY,GAAG,QAAQ,GAAG,kBAAkB,CAAC;IAC9D,6CAA6C;IAC7C,IAAI,EAAE,MAAM,CAAC;IACb,yDAAyD;IACzD,EAAE,CAAC,EAAE,MAAM,CAAC;IACZ,6DAA6D;IAC7D,WAAW,CAAC,EAAE,MAAM,CAAC;IACrB;;;;;OAKG;IACH,MAAM,CAAC,EAAE,MAAM,CAAC;IAChB,kFAAkF;IAClF,cAAc,CAAC,EAAE,MAAM,EAAE,CAAC;CAC3B;AAED;;;;;;;GAOG;AACH,MAAM,WAAW,kBAAkB;IACjC,0EAA0E;IAC1E,IAAI,EAAE,MAAM,CAAC;IACb;;;;OAIG;IACH,QAAQ,EAAE,MAAM,CAAC;IACjB;;;;OAIG;IACH,cAAc,EAAE,MAAM,CAAC;CACxB;AAED;;GAEG;AACH,MAAM,WAAW,iBAAiB;IAChC,gDAAgD;IAChD,OAAO,EAAE,MAAM,CAAC;IAChB,4DAA4D;IAC5D,WAAW,CAAC,EAAE,MAAM,CAAC;IACrB,gCAAgC;IAChC,UAAU,EAAE,aAAa,EAAE,CAAC;IAC5B,8CAA8C;IAC9C,SAAS,CAAC,EAAE,MAAM,CAAC;IACnB,qDAAqD;IACrD,QAAQ,CAAC,EAAE,OAAO,CAAC;IACnB,4EAA4E;IAC5E,gBAAgB,CAAC,EAAE,OAAO,CAAC;IAC3B,uEAAuE;IACvE,cAAc,CAAC,EAAE,MAAM,CAAC;IACxB,mEAAmE;IACnE,cAAc,CAAC,EAAE,MAAM,CAAC;IACxB;;;;OAIG;IACH,mBAAmB,CAAC,EAAE,kBAAkB,EAAE,CAAC;CAC5C;AAED;;GAEG;AACH,MAAM,MAAM,uBAAuB,GAC/B,MAAM,GACN,SAAS,GACT,UAAU,GACV,MAAM,CAAC;AAEX;;GAEG;AACH,MAAM,WAAW,uBAAwB,SAAQ,aAAa;IAC5D,cAAc,EAAE,uBAAuB,CAAC;CACzC;AAED;;GAEG;AACH,MAAM,WAAW,oBAAoB;IACnC,8CAA8C;IAC9C,IAAI,EAAE,aAAa,EAAE,CAAC;IACtB,8CAA8C;IAC9C,OAAO,EAAE,aAAa,EAAE,CAAC;IACzB,wCAAwC;IACxC,QAAQ,EAAE,aAAa,EAAE,CAAC;IAC1B,oCAAoC;IACpC,IAAI,EAAE,aAAa,EAAE,CAAC;CACvB;AAED;;GAEG;AACH,MAAM,WAAW,eAAe;IAC9B,8BAA8B;IAC9B,OAAO,EAAE,MAAM,CAAC;IAChB,8BAA8B;IAC9B,OAAO,EAAE,MAAM,CAAC;IAChB,gEAAgE;IAChE,OAAO,EAAE,MAAM,CAAC;IAChB,sCAAsC;IACtC,SAAS,EAAE,MAAM,CAAC;CACnB;AAED;;GAEG;AACH,MAAM,MAAM,eAAe,GACvB,QAAQ,GACR,eAAe,GACf,MAAM,GACN,WAAW,CAAC;AAEhB;;;GAGG;AACH,MAAM,MAAM,cAAc,GAAG,MAAM,CAAC,MAAM,EAAE,MAAM,CAAC,CAAC"}
+{"version":3,"file":"migration.d.ts","sourceRoot":"","sources":["../../src/types/migration.ts"],"names":[],"mappings":"AAAA;;;;;GAKG;AAEH;;GAEG;AACH,MAAM,WAAW,aAAa;IAC5B,+BAA+B;IAC/B,IAAI,EAAE,QAAQ,GAAG,YAAY,GAAG,QAAQ,GAAG,kBAAkB,CAAC;IAC9D,6CAA6C;IAC7C,IAAI,EAAE,MAAM,CAAC;IACb,yDAAyD;IACzD,EAAE,CAAC,EAAE,MAAM,CAAC;IACZ,6DAA6D;IAC7D,WAAW,CAAC,EAAE,MAAM,CAAC;IACrB;;;;;OAKG;IACH,MAAM,CAAC,EAAE,MAAM,CAAC;IAChB,kFAAkF;IAClF,cAAc,CAAC,EAAE,MAAM,EAAE,CAAC;CAC3B;AAED;;;;;;;GAOG;AACH,MAAM,WAAW,kBAAkB;IACjC,uEAAuE;IACvE,IAAI,EAAE,MAAM,CAAC;IACb;;;;OAIG;IACH,QAAQ,EAAE,MAAM,CAAC;IACjB;;;;OAIG;IACH,cAAc,EAAE,MAAM,CAAC;CACxB;AAED;;GAEG;AACH,MAAM,WAAW,iBAAiB;IAChC,gDAAgD;IAChD,OAAO,EAAE,MAAM,CAAC;IAChB,4DAA4D;IAC5D,WAAW,CAAC,EAAE,MAAM,CAAC;IACrB,gCAAgC;IAChC,UAAU,EAAE,aAAa,EAAE,CAAC;IAC5B,8CAA8C;IAC9C,SAAS,CAAC,EAAE,MAAM,CAAC;IACnB,qDAAqD;IACrD,QAAQ,CAAC,EAAE,OAAO,CAAC;IACnB,4EAA4E;IAC5E,gBAAgB,CAAC,EAAE,OAAO,CAAC;IAC3B,uEAAuE;IACvE,cAAc,CAAC,EAAE,MAAM,CAAC;IACxB,mEAAmE;IACnE,cAAc,CAAC,EAAE,MAAM,CAAC;IACxB;;;;OAIG;IACH,mBAAmB,CAAC,EAAE,kBAAkB,EAAE,CAAC;CAC5C;AAED;;GAEG;AACH,MAAM,MAAM,uBAAuB,GAC/B,MAAM,GACN,SAAS,GACT,UAAU,GACV,MAAM,CAAC;AAEX;;GAEG;AACH,MAAM,WAAW,uBAAwB,SAAQ,aAAa;IAC5D,cAAc,EAAE,uBAAuB,CAAC;CACzC;AAED;;GAEG;AACH,MAAM,WAAW,oBAAoB;IACnC,8CAA8C;IAC9C,IAAI,EAAE,aAAa,EAAE,CAAC;IACtB,8CAA8C;IAC9C,OAAO,EAAE,aAAa,EAAE,CAAC;IACzB,wCAAwC;IACxC,QAAQ,EAAE,aAAa,EAAE,CAAC;IAC1B,oCAAoC;IACpC,IAAI,EAAE,aAAa,EAAE,CAAC;CACvB;AAED;;GAEG;AACH,MAAM,WAAW,eAAe;IAC9B,8BAA8B;IAC9B,OAAO,EAAE,MAAM,CAAC;IAChB,8BAA8B;IAC9B,OAAO,EAAE,MAAM,CAAC;IAChB,gEAAgE;IAChE,OAAO,EAAE,MAAM,CAAC;IAChB,sCAAsC;IACtC,SAAS,EAAE,MAAM,CAAC;CACnB;AAED;;GAEG;AACH,MAAM,MAAM,eAAe,GACvB,QAAQ,GACR,eAAe,GACf,MAAM,GACN,WAAW,CAAC;AAEhB;;;GAGG;AACH,MAAM,MAAM,cAAc,GAAG,MAAM,CAAC,MAAM,EAAE,MAAM,CAAC,CAAC"}

package/dist/utils/agents-md.d.ts

@@ -0,0 +1,47 @@
+/**
+ * AGENTS.md managed-block helpers.
+ *
+ * cursor-trellis maintains a `<!-- CSTL:START -->...<!-- CSTL:END -->` block in
+ * AGENTS.md. Upstream mindfold-ai/Trellis maintains a separate
+ * `<!-- TRELLIS:START -->...<!-- TRELLIS:END -->` block. The two blocks can
+ * coexist in the same file so a repo can run upstream Trellis (other platforms)
+ * and cursor-trellis (Cursor) side by side.
+ *
+ * cursor-trellis never touches the upstream TRELLIS block or user content
+ * outside its own CSTL block.
+ */
+export declare const CSTL_BLOCK_START = "<!-- CSTL:START -->";
+export declare const CSTL_BLOCK_END = "<!-- CSTL:END -->";
+export declare const LEGACY_TRELLIS_BLOCK_START = "<!-- TRELLIS:START -->";
+export declare const LEGACY_TRELLIS_BLOCK_END = "<!-- TRELLIS:END -->";
+/** Extract the first block delimited by `startMarker`/`endMarker`, or null. */
+export declare function extractBlock(content: string, startMarker: string, endMarker: string): string | null;
+/**
+ * Insert or refresh the CSTL managed block in `existingContent` WITHOUT
+ * touching an upstream TRELLIS block or user content outside the blocks.
+ *
+ * - If `existingContent` already has a CSTL block → replace just that block.
+ * - Else if it has a legacy TRELLIS block → insert the CSTL block right after
+ *   the TRELLIS block (coexistence with upstream Trellis).
+ * - Else → append the CSTL block at the end.
+ *
+ * `templateContent` must contain a CSTL block; its block is extracted and used
+ * as the inserted/refreshed CSTL block content. If `templateContent` has no
+ * CSTL block, the whole `templateContent` is used as the block (fallback).
+ */
+export declare function insertCstlManagedBlock(existingContent: string, templateContent: string): string;
+/** True if `content` contains a CSTL managed block. */
+export declare function hasCstlBlock(content: string): boolean;
+/** True if `content` contains a legacy TRELLIS managed block. */
+export declare function hasLegacyTrellisBlock(content: string): boolean;
+/**
+ * Remove the CSTL managed block from `content`, leaving any upstream TRELLIS
+ * block and user content untouched. Collapses the blank lines left behind so
+ * the result reads naturally. Returns the content unchanged if no CSTL block
+ * is present.
+ *
+ * Used by `cstl uninstall` so a coexistence repo ends up with only the
+ * upstream TRELLIS block + user content after cursor-trellis is removed.
+ */
+export declare function removeCstlManagedBlock(content: string): string;
+//# sourceMappingURL=agents-md.d.ts.map

package/dist/utils/agents-md.d.ts.map

@@ -0,0 +1 @@
+{"version":3,"file":"agents-md.d.ts","sourceRoot":"","sources":["../../src/utils/agents-md.ts"],"names":[],"mappings":"AAAA;;;;;;;;;;;GAWG;AAEH,eAAO,MAAM,gBAAgB,wBAAwB,CAAC;AACtD,eAAO,MAAM,cAAc,sBAAsB,CAAC;AAClD,eAAO,MAAM,0BAA0B,2BAA2B,CAAC;AACnE,eAAO,MAAM,wBAAwB,yBAAyB,CAAC;AAE/D,+EAA+E;AAC/E,wBAAgB,YAAY,CAC1B,OAAO,EAAE,MAAM,EACf,WAAW,EAAE,MAAM,EACnB,SAAS,EAAE,MAAM,GAChB,MAAM,GAAG,IAAI,CAMf;AAED;;;;;;;;;;;;GAYG;AACH,wBAAgB,sBAAsB,CACpC,eAAe,EAAE,MAAM,EACvB,eAAe,EAAE,MAAM,GACtB,MAAM,CAwCR;AAED,uDAAuD;AACvD,wBAAgB,YAAY,CAAC,OAAO,EAAE,MAAM,GAAG,OAAO,CAMrD;AAED,iEAAiE;AACjE,wBAAgB,qBAAqB,CAAC,OAAO,EAAE,MAAM,GAAG,OAAO,CAM9D;AAED;;;;;;;;GAQG;AACH,wBAAgB,sBAAsB,CAAC,OAAO,EAAE,MAAM,GAAG,MAAM,CAiB9D"}

package/dist/utils/agents-md.js

@@ -0,0 +1,107 @@
+/**
+ * AGENTS.md managed-block helpers.
+ *
+ * cursor-trellis maintains a `<!-- CSTL:START -->...<!-- CSTL:END -->` block in
+ * AGENTS.md. Upstream mindfold-ai/Trellis maintains a separate
+ * `<!-- TRELLIS:START -->...<!-- TRELLIS:END -->` block. The two blocks can
+ * coexist in the same file so a repo can run upstream Trellis (other platforms)
+ * and cursor-trellis (Cursor) side by side.
+ *
+ * cursor-trellis never touches the upstream TRELLIS block or user content
+ * outside its own CSTL block.
+ */
+export const CSTL_BLOCK_START = "<!-- CSTL:START -->";
+export const CSTL_BLOCK_END = "<!-- CSTL:END -->";
+export const LEGACY_TRELLIS_BLOCK_START = "<!-- TRELLIS:START -->";
+export const LEGACY_TRELLIS_BLOCK_END = "<!-- TRELLIS:END -->";
+/** Extract the first block delimited by `startMarker`/`endMarker`, or null. */
+export function extractBlock(content, startMarker, endMarker) {
+    const start = content.indexOf(startMarker);
+    if (start === -1)
+        return null;
+    const end = content.indexOf(endMarker, start);
+    if (end === -1)
+        return null;
+    return content.slice(start, end + endMarker.length);
+}
+/**
+ * Insert or refresh the CSTL managed block in `existingContent` WITHOUT
+ * touching an upstream TRELLIS block or user content outside the blocks.
+ *
+ * - If `existingContent` already has a CSTL block → replace just that block.
+ * - Else if it has a legacy TRELLIS block → insert the CSTL block right after
+ *   the TRELLIS block (coexistence with upstream Trellis).
+ * - Else → append the CSTL block at the end.
+ *
+ * `templateContent` must contain a CSTL block; its block is extracted and used
+ * as the inserted/refreshed CSTL block content. If `templateContent` has no
+ * CSTL block, the whole `templateContent` is used as the block (fallback).
+ */
+export function insertCstlManagedBlock(existingContent, templateContent) {
+    const block = extractBlock(templateContent, CSTL_BLOCK_START, CSTL_BLOCK_END) ??
+        templateContent;
+    // Case 1: existing already has a CSTL block → replace it in place.
+    const cstlStart = existingContent.indexOf(CSTL_BLOCK_START);
+    if (cstlStart !== -1) {
+        const cstlEnd = existingContent.indexOf(CSTL_BLOCK_END, cstlStart);
+        if (cstlEnd !== -1) {
+            return (existingContent.slice(0, cstlStart) +
+                block +
+                existingContent.slice(cstlEnd + CSTL_BLOCK_END.length));
+        }
+    }
+    // Case 2: existing has a legacy TRELLIS block → coexist: insert CSTL block
+    // immediately after the TRELLIS block. Never modify the TRELLIS block.
+    const trellisStart = existingContent.indexOf(LEGACY_TRELLIS_BLOCK_START);
+    if (trellisStart !== -1) {
+        const trellisEnd = existingContent.indexOf(LEGACY_TRELLIS_BLOCK_END, trellisStart);
+        if (trellisEnd !== -1) {
+            const after = trellisEnd + LEGACY_TRELLIS_BLOCK_END.length;
+            return (existingContent.slice(0, after) +
+                "\n\n" +
+                block +
+                existingContent.slice(after));
+        }
+    }
+    // Case 3: no managed block → append at end, preserving any user content.
+    const trimmed = existingContent.replace(/\s+$/, "");
+    return `${trimmed}\n\n${block}\n`;
+}
+/** True if `content` contains a CSTL managed block. */
+export function hasCstlBlock(content) {
+    return (content.includes(CSTL_BLOCK_START) &&
+        content.indexOf(CSTL_BLOCK_END) >
+            content.indexOf(CSTL_BLOCK_START));
+}
+/** True if `content` contains a legacy TRELLIS managed block. */
+export function hasLegacyTrellisBlock(content) {
+    return (content.includes(LEGACY_TRELLIS_BLOCK_START) &&
+        content.indexOf(LEGACY_TRELLIS_BLOCK_END) >
+            content.indexOf(LEGACY_TRELLIS_BLOCK_START));
+}
+/**
+ * Remove the CSTL managed block from `content`, leaving any upstream TRELLIS
+ * block and user content untouched. Collapses the blank lines left behind so
+ * the result reads naturally. Returns the content unchanged if no CSTL block
+ * is present.
+ *
+ * Used by `cstl uninstall` so a coexistence repo ends up with only the
+ * upstream TRELLIS block + user content after cursor-trellis is removed.
+ */
+export function removeCstlManagedBlock(content) {
+    const startIdx = content.indexOf(CSTL_BLOCK_START);
+    if (startIdx === -1) {
+        return content;
+    }
+    const endIdx = content.indexOf(CSTL_BLOCK_END, startIdx);
+    if (endIdx === -1) {
+        return content;
+    }
+    const before = content.slice(0, startIdx);
+    const after = content.slice(endIdx + CSTL_BLOCK_END.length);
+    // Collapse up to one blank-line gap on each side so we don't leave a run of
+    // 4+ newlines, but never trim away meaningful separator lines.
+    return (before.replace(/\n{2,}$/, "\n\n") +
+        after.replace(/^\n{2,}/, "\n\n")).replace(/^\n{2,}/, "\n");
+}
+//# sourceMappingURL=agents-md.js.map

package/dist/utils/agents-md.js.map

@@ -0,0 +1 @@
+{"version":3,"file":"agents-md.js","sourceRoot":"","sources":["../../src/utils/agents-md.ts"],"names":[],"mappings":"AAAA;;;;;;;;;;;GAWG;AAEH,MAAM,CAAC,MAAM,gBAAgB,GAAG,qBAAqB,CAAC;AACtD,MAAM,CAAC,MAAM,cAAc,GAAG,mBAAmB,CAAC;AAClD,MAAM,CAAC,MAAM,0BAA0B,GAAG,wBAAwB,CAAC;AACnE,MAAM,CAAC,MAAM,wBAAwB,GAAG,sBAAsB,CAAC;AAE/D,+EAA+E;AAC/E,MAAM,UAAU,YAAY,CAC1B,OAAe,EACf,WAAmB,EACnB,SAAiB;IAEjB,MAAM,KAAK,GAAG,OAAO,CAAC,OAAO,CAAC,WAAW,CAAC,CAAC;IAC3C,IAAI,KAAK,KAAK,CAAC,CAAC;QAAE,OAAO,IAAI,CAAC;IAC9B,MAAM,GAAG,GAAG,OAAO,CAAC,OAAO,CAAC,SAAS,EAAE,KAAK,CAAC,CAAC;IAC9C,IAAI,GAAG,KAAK,CAAC,CAAC;QAAE,OAAO,IAAI,CAAC;IAC5B,OAAO,OAAO,CAAC,KAAK,CAAC,KAAK,EAAE,GAAG,GAAG,SAAS,CAAC,MAAM,CAAC,CAAC;AACtD,CAAC;AAED;;;;;;;;;;;;GAYG;AACH,MAAM,UAAU,sBAAsB,CACpC,eAAuB,EACvB,eAAuB;IAEvB,MAAM,KAAK,GACT,YAAY,CAAC,eAAe,EAAE,gBAAgB,EAAE,cAAc,CAAC;QAC/D,eAAe,CAAC;IAElB,mEAAmE;IACnE,MAAM,SAAS,GAAG,eAAe,CAAC,OAAO,CAAC,gBAAgB,CAAC,CAAC;IAC5D,IAAI,SAAS,KAAK,CAAC,CAAC,EAAE,CAAC;QACrB,MAAM,OAAO,GAAG,eAAe,CAAC,OAAO,CAAC,cAAc,EAAE,SAAS,CAAC,CAAC;QACnE,IAAI,OAAO,KAAK,CAAC,CAAC,EAAE,CAAC;YACnB,OAAO,CACL,eAAe,CAAC,KAAK,CAAC,CAAC,EAAE,SAAS,CAAC;gBACnC,KAAK;gBACL,eAAe,CAAC,KAAK,CAAC,OAAO,GAAG,cAAc,CAAC,MAAM,CAAC,CACvD,CAAC;QACJ,CAAC;IACH,CAAC;IAED,2EAA2E;IAC3E,uEAAuE;IACvE,MAAM,YAAY,GAAG,eAAe,CAAC,OAAO,CAAC,0BAA0B,CAAC,CAAC;IACzE,IAAI,YAAY,KAAK,CAAC,CAAC,EAAE,CAAC;QACxB,MAAM,UAAU,GAAG,eAAe,CAAC,OAAO,CACxC,wBAAwB,EACxB,YAAY,CACb,CAAC;QACF,IAAI,UAAU,KAAK,CAAC,CAAC,EAAE,CAAC;YACtB,MAAM,KAAK,GAAG,UAAU,GAAG,wBAAwB,CAAC,MAAM,CAAC;YAC3D,OAAO,CACL,eAAe,CAAC,KAAK,CAAC,CAAC,EAAE,KAAK,CAAC;gBAC/B,MAAM;gBACN,KAAK;gBACL,eAAe,CAAC,KAAK,CAAC,KAAK,CAAC,CAC7B,CAAC;QACJ,CAAC;IACH,CAAC;IAED,yEAAyE;IACzE,MAAM,OAAO,GAAG,eAAe,CAAC,OAAO,CAAC,MAAM,EAAE,EAAE,CAAC,CAAC;IACpD,OAAO,GAAG,OAAO,OAAO,KAAK,IAAI,CAAC;AACpC,CAAC;AAED,uDAAuD;AACvD,MAAM,UAAU,YAAY,CAAC,OAAe;IAC1C,OAAO,CACL,OAAO,CAAC,QAAQ,CAAC,gBAAgB,CAAC;QAClC,OAAO,CAAC,OAAO,CAAC,cAAc,CAAC;YAC7B,OAAO,CAAC,OAAO,CAAC,gBAAgB,CAAC,CACpC,CAAC;AACJ,CAAC;AAED,iEAAiE;AACjE,MAAM,UAAU,qBAAqB,CAAC,OAAe;IACnD,OAAO,CACL,OAAO,CAAC,QAAQ,CAAC,0BAA0B,CAAC;QAC5C,OAAO,CAAC,OAAO,CAAC,wBAAwB,CAAC;YACvC,OAAO,CAAC,OAAO,CAAC,0BAA0B,CAAC,CAC9C,CAAC;AACJ,CAAC;AAED;;;;;;;;GAQG;AACH,MAAM,UAAU,sBAAsB,CAAC,OAAe;IACpD,MAAM,QAAQ,GAAG,OAAO,CAAC,OAAO,CAAC,gBAAgB,CAAC,CAAC;IACnD,IAAI,QAAQ,KAAK,CAAC,CAAC,EAAE,CAAC;QACpB,OAAO,OAAO,CAAC;IACjB,CAAC;IACD,MAAM,MAAM,GAAG,OAAO,CAAC,OAAO,CAAC,cAAc,EAAE,QAAQ,CAAC,CAAC;IACzD,IAAI,MAAM,KAAK,CAAC,CAAC,EAAE,CAAC;QAClB,OAAO,OAAO,CAAC;IACjB,CAAC;IACD,MAAM,MAAM,GAAG,OAAO,CAAC,KAAK,CAAC,CAAC,EAAE,QAAQ,CAAC,CAAC;IAC1C,MAAM,KAAK,GAAG,OAAO,CAAC,KAAK,CAAC,MAAM,GAAG,cAAc,CAAC,MAAM,CAAC,CAAC;IAC5D,4EAA4E;IAC5E,+DAA+D;IAC/D,OAAO,CACL,MAAM,CAAC,OAAO,CAAC,SAAS,EAAE,MAAM,CAAC;QACjC,KAAK,CAAC,OAAO,CAAC,SAAS,EAAE,MAAM,CAAC,CACjC,CAAC,OAAO,CAAC,SAAS,EAAE,IAAI,CAAC,CAAC;AAC7B,CAAC"}

package/dist/utils/codebase-retrieval-router.js

@@ -187,7 +187,7 @@ function policyVerification() {
     return [
         {
             id: "policy-doc-top1",
-            requirement: "For policy/document intents, confirm Top-1 policy evidence from AGENTS.md or .trellis/spec before ranking implementation modules first.",
+            requirement: "For policy/document intents, confirm Top-1 policy evidence from AGENTS.md or .cstl/spec before ranking implementation modules first.",
             appliesToRoles: ["exact", "semantic"],
         },
         {
@@ -345,7 +345,7 @@ function orderedRoutesForIntents(intents, includeOptionalAdapters, projectFileCo
                 role: "exact",
                 sourceFamily: "policy-docs",
                 commands: [
-                    'rg -i "storage default|sidecar|sqlite only" AGENTS.md "**/AGENTS.md" README.md CONTRIBUTING.md .trellis/spec',
+                    'rg -i "storage default|sidecar|sqlite only" AGENTS.md "**/AGENTS.md" README.md CONTRIBUTING.md .cstl/spec',
                 ],
                 rationale: "Policy/document intent: search instruction and spec docs first.",
             });
@@ -369,7 +369,7 @@ function orderedRoutesForIntents(intents, includeOptionalAdapters, projectFileCo
             role: "exact",
             sourceFamily: "policy-docs",
             commands: [
-                'rg -i "storage default|sidecar|sqlite only" AGENTS.md "**/AGENTS.md" README.md CONTRIBUTING.md .trellis/spec',
+                'rg -i "storage default|sidecar|sqlite only" AGENTS.md "**/AGENTS.md" README.md CONTRIBUTING.md .cstl/spec',
             ],
             rationale: "Policy/document intent: search instruction and spec docs first.",
         });
@@ -395,7 +395,7 @@ function orderedRoutesForIntents(intents, includeOptionalAdapters, projectFileCo
                 role: "exact",
                 sourceFamily: "policy-docs",
                 commands: [
-                    'rg -i "storage default|sidecar|sqlite only" AGENTS.md "**/AGENTS.md" README.md CONTRIBUTING.md .trellis/spec',
+                    'rg -i "storage default|sidecar|sqlite only" AGENTS.md "**/AGENTS.md" README.md CONTRIBUTING.md .cstl/spec',
                 ],
                 rationale: "Policy/document branch after exact-primary when both intents match.",
             });
@@ -547,7 +547,7 @@ function buildFallbackHints(intents, includeOptionalAdapters, routes, cursorEnv
     if (intents.some((i) => i.id === "policy-document")) {
         hints.push({
             when: "semantic Top-1 is implementation-only for policy query",
-            action: "Fall back to policy-doc rg route and AGENTS.md/.trellis/spec reads before accepting semantic ranking.",
+            action: "Fall back to policy-doc rg route and AGENTS.md/.cstl/spec reads before accepting semantic ranking.",
             replacesRole: "semantic",
         });
     }