docs: Cursor MCP live smoke evidence (traverse-mcp Mode A)

apps/llm-mcp-reference/clients/cursor/README.md

@@ -2,11 +2,17 @@
 
 ## Setup
 
-1. Traverse checkout with working `traverse-mcp`.
-2. Copy [`mcp.json.example`](mcp.json.example) into Cursor MCP settings (or project `.cursor`/MCP config as applicable).
-3. Fix absolute paths.
+1. Traverse checkout with working `traverse-mcp` (tested with **v0.8.2+**).
+2. Copy [`mcp.json.example`](mcp.json.example) into Cursor MCP settings (Cursor Settings → MCP, or project MCP config as applicable).
+3. Replace `/ABS/PATH/TO/Traverse` with your **absolute** Traverse checkout in both `cwd` and `env.TRAVERSE_REPO`.
 4. Apply [`../../shared/prompts/system-boundary.md`](../../shared/prompts/system-boundary.md) in rules/instructions.
-5. Exercise [`../../shared/workflows/traverse-starter.md`](../../shared/workflows/traverse-starter.md).
-6. For meeting transcripts, follow [`../../shared/workflows/meeting-notes.md`](../../shared/workflows/meeting-notes.md) (tool sequence + sample transcript).
+5. Exercise [`../../shared/workflows/traverse-starter.md`](../../shared/workflows/traverse-starter.md) when starter entrypoints are available in the MCP catalog; until then use discovery (`list_entrypoints`) and only render runtime fields.
+6. For meeting transcripts, follow [`../../shared/workflows/meeting-notes.md`](../../shared/workflows/meeting-notes.md) (tool sequence + sample transcript) once `meeting-notes.process` is catalogued.
+
+## Live evidence
+
+See [`evidence/`](evidence/) — Cursor cloud agent ran the **same** `cargo run -p traverse-mcp -- stdio` command as `mcp.json.example` and captured discovery + execute/render transcripts (`llm-mcp-cursor-live-smoke`).
+
+## Boundary
 
 Cursor agents working **in this repo** still claim OS tickets via `AGENTS.md`; this façade is for **product** workflows through Traverse MCP, not for replacing Project 2 claim locks.

apps/llm-mcp-reference/clients/cursor/evidence/README.md

@@ -0,0 +1,56 @@
+# Cursor MCP live evidence (`llm-mcp-cursor-live-smoke`)
+
+Captured **2026-07-29** against Traverse **v0.8.2** (`TRAVERSE_REPO=/tmp/Traverse`).
+
+## What was proven
+
+The Cursor façade config in [`../mcp.json.example`](../mcp.json.example) launches:
+
+```bash
+cargo run -p traverse-mcp -- stdio
+```
+
+with `cwd` / `TRAVERSE_REPO` pointing at a Traverse checkout. That exact command was exercised here (stdio JSONL envelopes, local_trust Mode A).
+
+### Transcript A — discovery
+
+File: [`cursor-mcp-stdio-transcript.jsonl`](cursor-mcp-stdio-transcript.jsonl)
+
+Commands: `describe_server` → `list_entrypoints` → `list_content_groups` → `shutdown`
+
+Observed kinds: `mcp_stdio_server_startup` (ready), `mcp_stdio_server_description`, `mcp_stdio_server_entrypoint_list`, `mcp_stdio_server_content_group_list`, `mcp_stdio_server_shutdown`.
+
+### transcript B — validate / execute / render
+
+File: [`cursor-mcp-execute-transcript.jsonl`](cursor-mcp-execute-transcript.jsonl)
+
+Commands follow the upstream `mcp_stdio_server_execution_report_smoke.sh` path against the **current** MCP catalog workflow `expedition.planning.plan-expedition` (request_path `examples/expedition/runtime-requests/plan-expedition.json`).
+
+Observed: validation `valid` → execution `completed` → report `rendered` → shutdown `complete`. stderr empty.
+
+## Runtime fields only
+
+Execution/report envelopes come from Traverse MCP — this façade does **not** invent structured business fields. Agents must present only fields returned by `render_execution_report` / execution envelopes.
+
+## Catalog note (honest)
+
+As of Traverse v0.8.2, `traverse-mcp stdio`’s default governed catalog is the **expedition** registry bundle (plus `core-runtime-example` content group). **`traverse-starter.*` / `meeting-notes.process` entrypoints are not listed yet** in that default catalog.
+
+Implication for App-Refs:
+
+- Cursor MCP config is correct for Mode A (stdio → `traverse-mcp`).
+- Live product smoke against starter/meeting-notes **IDs** lands when those capabilities are published into the MCP catalog (or a documented catalog override ships upstream). Until then, this evidence proves the Cursor→MCP host path end-to-end on the catalog the server actually serves.
+
+## Reproduce
+
+```bash
+export TRAVERSE_REPO=/absolute/path/to/Traverse
+cd "$TRAVERSE_REPO"
+printf '%s\n' \
+  '{"command":"describe_server"}' \
+  '{"command":"list_entrypoints"}' \
+  '{"command":"shutdown"}' \
+  | cargo run -q -p traverse-mcp -- stdio
+```
+
+Optional full execute path: see Traverse `scripts/ci/mcp_stdio_server_execution_report_smoke.sh`.

apps/llm-mcp-reference/clients/cursor/evidence/cursor-mcp-stdio-transcript.jsonl

@@ -0,0 +1,5 @@
+{"auth_boundary":{"default_mode":"local_trust","required_commands":["execute_entrypoint","render_execution_report"],"required_token_env":"TRAVERSE_MCP_STDIO_BEARER_TOKEN"},"content_group_count":1,"governing_spec":"022-mcp-wasm-server","host_mode":"stdio","kind":"mcp_stdio_server_startup","public_surface_id":"traverse.mcp.stdio-server","server_name":"traverse-mcp","status":"ready","supported_commands":["describe_server","list_content_groups","describe_content_group","list_entrypoints","describe_entrypoint","validate_entrypoint","execute_entrypoint","render_execution_report","shutdown"]}
+{"auth_boundary":{"default_mode":"local_trust","required_commands":["execute_entrypoint","render_execution_report"],"required_token_env":"TRAVERSE_MCP_STDIO_BEARER_TOKEN","token_output_policy":"never_echo_raw_token"},"content_groups":[{"content_group_id":"core-runtime-example","display_name":"Core runtime example","governed_paths":["examples/templates/executable-capability-package/manifest.template.json","docs/executable-package-template.md","docs/local-runtime-home.md","scripts/ci/executable_package_template_smoke.sh"],"invocable_entrypoints":["describe_content_group"],"summary":"Traverse-neutral executable capability package template and local runtime shape.","validation_commands":["bash scripts/ci/executable_package_template_smoke.sh"]}],"downstream_validation_path":{"consumer_name":"youaskm3","governing_specs":["019-downstream-consumer-contract","020-downstream-integration-validation","021-app-facing-operational-constraints"],"public_surface_id":"traverse.mcp.downstream-consumer","validated_flow_id":"youaskm3_mcp_validation"},"governed_surface_counts":{"capabilities":6,"events":5,"workflows":1},"governing_spec":"022-mcp-wasm-server","host_mode":"stdio","kind":"mcp_stdio_server_description","public_surface_id":"traverse.mcp.stdio-server","runtime_authority":"Traverse runtime authority","server_name":"traverse-mcp","supported_commands":["describe_server","list_content_groups","describe_content_group","list_entrypoints","describe_entrypoint","validate_entrypoint","execute_entrypoint","render_execution_report","shutdown"]}
+{"content_groups":[{"content_group_id":"core-runtime-example","display_name":"Core runtime example","governed_paths":["examples/templates/executable-capability-package/manifest.template.json","docs/executable-package-template.md","docs/local-runtime-home.md","scripts/ci/executable_package_template_smoke.sh"],"invocable_entrypoints":["describe_content_group"],"summary":"Traverse-neutral executable capability package template and local runtime shape.","validation_commands":["bash scripts/ci/executable_package_template_smoke.sh"]}],"entrypoints":{"capabilities":[{"artifact_kind":"capability","id":"expedition.planning.capture-expedition-objective","lifecycle":"active","summary":"Capture a structured expedition objective from destination, timing, and planning preference inputs.","version":"1.0.0"},{"artifact_kind":"capability","id":"expedition.planning.interpret-expedition-intent","lifecycle":"active","summary":"Interpret free-form expedition planning intent into structured route preferences, constraints, and assumptions.","version":"1.0.0"},{"artifact_kind":"capability","id":"expedition.planning.assess-conditions-summary","lifecycle":"active","summary":"Assess a deterministic expedition conditions summary from objective and interpreted intent context.","version":"1.0.0"},{"artifact_kind":"capability","id":"expedition.planning.validate-team-readiness","lifecycle":"active","summary":"Validate expedition team readiness from objective, conditions summary, and team profile context.","version":"1.0.0"},{"artifact_kind":"capability","id":"expedition.planning.assemble-expedition-plan","lifecycle":"active","summary":"Assemble the final expedition plan from objective, interpreted intent, conditions, and readiness inputs.","version":"1.0.0"},{"artifact_kind":"capability","id":"expedition.planning.plan-expedition","lifecycle":"active","summary":"Compose the canonical expedition planning workflow into one first-class governed capability.","version":"1.0.0"}],"events":[{"artifact_kind":"event","id":"expedition.planning.expedition-objective-captured","lifecycle":"active","summary":"A normalized expedition objective has been captured and is ready for downstream planning.","version":"1.0.0"},{"artifact_kind":"event","id":"expedition.planning.expedition-intent-interpreted","lifecycle":"active","summary":"An expedition intent has been interpreted into structured planning inputs for downstream steps.","version":"1.0.0"},{"artifact_kind":"event","id":"expedition.planning.conditions-summary-assessed","lifecycle":"active","summary":"A deterministic expedition conditions summary has been assessed for the current planning context.","version":"1.0.0"},{"artifact_kind":"event","id":"expedition.planning.team-readiness-validated","lifecycle":"active","summary":"A team readiness result has been validated for the expedition objective and conditions context.","version":"1.0.0"},{"artifact_kind":"event","id":"expedition.planning.expedition-plan-assembled","lifecycle":"active","summary":"A final expedition plan has been assembled and is ready for downstream consumption.","version":"1.0.0"}],"workflows":[{"artifact_kind":"workflow","id":"expedition.planning.plan-expedition","lifecycle":"active","summary":"Capture, interpret, validate, and assemble an expedition plan deterministically.","version":"1.0.0"}]},"governing_spec":"022-mcp-wasm-server","host_mode":"stdio","kind":"mcp_stdio_server_entrypoint_list","server_name":"traverse-mcp"}
+{"content_groups":[{"content_group_id":"core-runtime-example","display_name":"Core runtime example","governed_paths":["examples/templates/executable-capability-package/manifest.template.json","docs/executable-package-template.md","docs/local-runtime-home.md","scripts/ci/executable_package_template_smoke.sh"],"invocable_entrypoints":["describe_content_group"],"summary":"Traverse-neutral executable capability package template and local runtime shape.","validation_commands":["bash scripts/ci/executable_package_template_smoke.sh"]}],"governing_spec":"022-mcp-wasm-server","host_mode":"stdio","kind":"mcp_stdio_server_content_group_list","server_name":"traverse-mcp"}
+{"governing_spec":"022-mcp-wasm-server","host_mode":"stdio","kind":"mcp_stdio_server_shutdown","reason":"shutdown_command","server_name":"traverse-mcp","status":"complete"}

docs/llm-reference-apps-plan.md

@@ -87,8 +87,8 @@ v1 documents **Mode A** only. Do not revive HTTP `traverse-cli serve` as the pro
 |---|---|---|
 | `llm-mcp-reference-apps-plan` | Plan + scaffold (this doc + tree) | Done (#238) |
 | `llm-mcp-claude-live-smoke` | Live Claude Desktop/Code path against `traverse-starter` | Ready |
-| `llm-mcp-cursor-live-smoke` | Live Cursor MCP path | Ready |
-| `llm-mcp-meeting-notes-workflow` | Document + config for meeting-notes via MCP | In Progress |
+| `llm-mcp-cursor-live-smoke` | Live Cursor MCP path | In Progress |
+| `llm-mcp-meeting-notes-workflow` | Document + config for meeting-notes via MCP | Done (#239) |
 | `llm-mcp-chatgpt-adapter` | ChatGPT Actions/GPT mapping (or MCP when shipped) | Future / Ready when API stable |
 | `llm-mcp-grok-adapter` | Grok tool-calling mapping | Future / Ready when API stable |
 | `llm-mcp-embedded-host` | Mode B embedded prepare/cache for MCP host | Blocked on Traverse Spec 520 implement |

scripts/ci/llm_mcp_reference_smoke.sh

@@ -22,6 +22,9 @@ need apps/llm-mcp-reference/clients/claude-code/README.md
 need apps/llm-mcp-reference/clients/claude-code/.mcp.json.example
 need apps/llm-mcp-reference/clients/cursor/README.md
 need apps/llm-mcp-reference/clients/cursor/mcp.json.example
+need apps/llm-mcp-reference/clients/cursor/evidence/README.md
+need apps/llm-mcp-reference/clients/cursor/evidence/cursor-mcp-stdio-transcript.jsonl
+need apps/llm-mcp-reference/clients/cursor/evidence/cursor-mcp-execute-transcript.jsonl
 need apps/llm-mcp-reference/clients/chatgpt/README.md
 need apps/llm-mcp-reference/clients/grok/README.md
 # Configs must mention traverse-mcp, not invent business fields