Update Cargo.lock with new dependencies: adler2 v2.0.1, crc32fast v1.5.0, flate2 v1.1.9, miniz_oxide v0.8.9, simd-adler32 v0.3.8, ureq v2.12.1, webpki-roots v0.26.11. Add related dependency entries and update existing ones.

Author: nikivdev

Cargo.lock

@@ -128,8 +128,11 @@ Key points:
   - `seq_*` tool definitions injected into new sessions
   - tool-name normalization (`seq_open_app`, `seq.open_app`, `seq:open-app`)
   - request correlation IDs (`request_id`, `run_id`, `tool_call_id`)
+- Event transport is SSE-first (`/sse`) with automatic fallback to polling (`/events`) if SSE is unavailable.
+- Optional Maple telemetry export can dual-write runtime traces to local + hosted ingest endpoints when `SEQ_EVERRUNS_MAPLE_*` env vars are configured.
 - Existing Flow features remain unchanged (`f seq-rpc`, session resume/copy/context flows).
 
 Setup and validation details are documented in:
 
 - `docs/everruns-seq-bridge-integration.md`
+- `docs/everruns-maple-runbook.md`

docs/commands/ai.md

@@ -69,6 +69,23 @@ f domains --engine native doctor
 - myflow-specific setup: `docs/myflow-localhost-runbook.md`.
 - Lifecycle integration: configure `[lifecycle.domains]` and use `f up` / `f down`.
 
+### macOS no-docker bind to :80
+
+If native startup fails with `Permission denied` on `127.0.0.1:80`, install launchd socket mode once:
+
+```bash
+cd ~/code/flow
+sudo ./tools/domainsd-cpp/install-macos-launchd.sh
+```
+
+Then run:
+
+```bash
+f domains --engine native up
+```
+
+This keeps routing fully native and avoids Docker overhead while still using port `80`.
+
 ### Native tuning envs
 
 You can tune the native daemon at startup via environment variables:

docs/commands/domains.md

@@ -19,6 +19,8 @@ f down
 - If `[lifecycle.domains]` is configured, optional teardown is applied:
   - `remove_on_down = true` -> removes configured host route
   - `stop_proxy_on_down = true` -> stops shared local domains proxy
+- On macOS launchd-managed native domains, stopping native proxy is handled by:
+  - `sudo ./tools/domainsd-cpp/uninstall-macos-launchd.sh`
 
 If neither a down task nor lifecycle domain teardown is configured, command fails with guidance.
 

docs/commands/down.md

@@ -0,0 +1,96 @@
+# Everruns + Maple Runbook
+
+This is the fastest path to use the new Everruns telemetry export now.
+
+It sends `f ai everruns` traces to:
+
+- local Maple (dev visualization)
+- hosted Maple (shared/history visualization)
+
+## What gets exported
+
+When enabled, Flow exports:
+
+- `everruns.tool_call` spans for each `seq_*` tool execution
+- runtime spans such as:
+  - `everruns.tool_call_requested`
+  - `everruns.output_message_completed`
+  - `everruns.turn_failed`
+
+## Prerequisites
+
+1. `seqd` is running and reachable at your socket (`/tmp/seqd.sock` by default).
+2. Everruns API is reachable (`http://127.0.0.1:9300/api` by default).
+3. You have Maple ingest keys for local and/or hosted endpoint.
+
+## 1) Configure env (now)
+
+From `~/code/flow`, set the endpoints + keys:
+
+```bash
+f env set SEQ_EVERRUNS_MAPLE_LOCAL_ENDPOINT=http://ingest.maple.localhost/v1/traces
+f env set SEQ_EVERRUNS_MAPLE_LOCAL_INGEST_KEY=maple_pk_local_xxx
+f env set SEQ_EVERRUNS_MAPLE_HOSTED_ENDPOINT=https://ingest.1focus.ai/v1/traces
+f env set SEQ_EVERRUNS_MAPLE_HOSTED_INGEST_KEY=maple_pk_hosted_xxx
+```
+
+Optional tuning:
+
+```bash
+f env set SEQ_EVERRUNS_MAPLE_QUEUE_CAPACITY=4096
+f env set SEQ_EVERRUNS_MAPLE_MAX_BATCH_SIZE=128
+f env set SEQ_EVERRUNS_MAPLE_FLUSH_INTERVAL_MS=50
+f env set SEQ_EVERRUNS_MAPLE_CONNECT_TIMEOUT_MS=400
+f env set SEQ_EVERRUNS_MAPLE_REQUEST_TIMEOUT_MS=800
+```
+
+For optimized mirror (remote ClickHouse + durable local spool), also set:
+
+```bash
+f env set SEQ_CH_MODE=mirror
+f env set SEQ_CH_MEM_PATH=~/repos/ClickHouse/ClickHouse/user_files/seq_mem.jsonl
+f env set SEQ_CH_LOG_PATH=~/repos/ClickHouse/ClickHouse/user_files/seq_trace.jsonl
+```
+
+## 2) Run with env injected
+
+Use `f env run` so runtime sees configured values:
+
+```bash
+f env run -- f ai everruns "open Safari and take a screenshot"
+```
+
+If you already export envs another way, this also works:
+
+```bash
+f ai everruns "open Safari and take a screenshot"
+```
+
+On startup, if telemetry is enabled, Flow prints:
+
+`maple dual-ingest telemetry enabled`
+
+## 3) Verify in Maple
+
+In Maple (local and hosted), filter by:
+
+- `service.name = seq-everruns-bridge`
+
+Look for span names:
+
+- `everruns.tool_call`
+- `everruns.tool_call_requested`
+- `everruns.output_message_completed`
+
+## Troubleshooting
+
+1. Error: `invalid SEQ_EVERRUNS_MAPLE_* configuration`
+   - You set only endpoint or only key for local/hosted pair.
+   - Fix by setting both or removing both for that pair.
+2. Everruns command works but no spans in Maple
+   - Confirm ingest endpoint includes `/v1/traces`.
+   - Confirm ingest key is valid for that endpoint.
+   - Confirm you ran through `f env run -- ...` (or equivalent env injection).
+3. Temporary Maple outage
+   - Tool execution continues.
+   - Export is best-effort and non-blocking.

docs/everruns-maple-runbook.md

@@ -26,6 +26,14 @@ Flow still owns and keeps unchanged:
 - Flow config/env resolution for Everruns (`[everruns]`, `FLOW_EVERRUNS_*`)
 - `f seq-rpc` command and other AI session commands
 
+Flow now additionally supports Maple dual-ingest telemetry export from the Everruns runtime
+when `SEQ_EVERRUNS_MAPLE_*` env vars are set.
+
+Runtime path is now SSE-first for lower latency:
+
+- primary: `GET /v1/sessions/{id}/sse` (push events, reconnect with `since_id`)
+- fallback: `GET /v1/sessions/{id}/events` polling when SSE endpoint is unavailable
+
 ## No-Overlap Contract
 
 This integration is intentionally scoped to avoid feature overlap:
@@ -34,6 +42,26 @@ This integration is intentionally scoped to avoid feature overlap:
 - Flow does not add a second Everruns runtime.
 - Existing `f ai claude` / `f ai codex` / `f seq-rpc` behavior remains unchanged.
 
+## Maple Dual Ingest (Local + Hosted)
+
+When enabled, Flow emits:
+
+- tool call spans (`everruns.tool_call`)
+- runtime event spans (`everruns.tool_call_requested`, `everruns.output_message_completed`, etc.)
+
+using `seq_everruns_bridge::maple::MapleTraceExporter` and non-blocking background batching.
+
+Required env keys:
+
+- `SEQ_EVERRUNS_MAPLE_LOCAL_ENDPOINT` (example: `http://ingest.maple.localhost/v1/traces`)
+- `SEQ_EVERRUNS_MAPLE_LOCAL_INGEST_KEY`
+- `SEQ_EVERRUNS_MAPLE_HOSTED_ENDPOINT` (example: `https://ingest.1focus.ai/v1/traces`)
+- `SEQ_EVERRUNS_MAPLE_HOSTED_INGEST_KEY`
+
+Operational setup/run instructions:
+
+- `docs/everruns-maple-runbook.md`
+
 ## Dependency Setup
 
 Current local setup in `Cargo.toml`:
@@ -67,6 +95,7 @@ cargo check
 cargo run --release --bin f -- ai everruns --help
 rg "bridge_tool_definitions|parse_tool_call_requested|bridge_build_request" src/ai_everruns.rs
 rg "map_seq_operation|seq_client_tool_definitions" src/ai_everruns.rs
+rg "MapleTraceExporter|MapleSpan::for_runtime_event" src/ai_everruns.rs
 ```
 
 End-to-end smoke (requires local Everruns + seqd):
@@ -79,7 +108,9 @@ Expected evidence of integration:
 
 - successful compile with shared bridge dependency
 - bridge call-sites present in Flow Everruns runtime
+- Maple exporter call-sites present (tool + runtime spans)
 - no duplicated tool catalog/mapping in `src/ai_everruns.rs`
+- SSE-first event consumption active in `src/ai_everruns.rs`
 
 ## When To Keep It / When To Revert
 

docs/everruns-seq-bridge-integration.md

@@ -27,6 +27,7 @@ High-signal docs for the Codex/Claude workflow:
 - [`myflow-localhost-runbook.md`](myflow-localhost-runbook.md): exact commands to run `~/code/myflow` on `myflow.localhost` and use in-app `/processes` and `/logs` views.
 - [`seq-agent-rpc-contract.md`](seq-agent-rpc-contract.md): hard interface contract for agent OS actions through `seq_client -> seqd` RPC v1.
 - [`everruns-seq-bridge-integration.md`](everruns-seq-bridge-integration.md): why/how `f ai everruns` now reuses Seq's shared Everruns bridge to avoid duplicate tool mapping logic.
+- [`everruns-maple-runbook.md`](everruns-maple-runbook.md): exact commands to enable Everruns -> Maple dual-ingest and verify spans in local + hosted Maple.
 - [`ascii-commit-visualization-pipeline.md`](ascii-commit-visualization-pipeline.md): end-to-end pipeline for commit analysis -> Flow API -> `box-of-rain` ASCII/SVG diagrams in myflow.
 - [`features.md`](features.md): practical command flows and examples.
 - [`moonbit-ai-tasks-implementation.md`](moonbit-ai-tasks-implementation.md): complete inventory of the task-centric MoonBit AI task runtime and Flow-local `.ai/tasks` pack.