Generated — do not edit by hand. This file is rendered from
roadmap.yamlbyscripts/gen-roadmap.py. Editroadmap.yamland re-run the generator.
The roadmap models cross-spec epics → tasks with a dependency DAG, execution status, priority, and links — the things a per-spec tasks.md checkbox list cannot express. Per-spec checkbox progress is recomputed live from each specs/<NNN>/tasks.md.
roadmap.yaml holds the working set (todo · in-progress · in-review · blocked · parked). Cold shipped epics are swept into roadmap.archive.yaml and surface in the Shipped table below, so the working file stays small while provenance survives. A depends_on: edge into the archive is satisfied by definition.
python3 scripts/gen-roadmap.py # writes ROADMAP.md
scripts/gen-roadmap # convenience wrapper (same thing)
python3 scripts/gen-roadmap.py --check # CI canary: fail if ROADMAP.md is stale
python3 scripts/gen-roadmap.py --check-github # cross-check statuses vs live GitHub PR state,
# spec links, depends_on ids, and status sanity
# (add --strict to fail on warnings; needs gh)
python3 scripts/gen-roadmap.py --archive --dry-run # preview the cold-done sweep
python3 scripts/gen-roadmap.py --archive # sweep into roadmap.archive.yaml + regenerate- epics[] — each has
id(stable slug, DAG node),title,status(todo·in_progress·in_review·blocked·done),priority(P0–P3),depends_on: [ids](DAG edges, prerequisite→dependent), optionalparked: true, and linksspec:/pr:/mcp:(external MCP-xxxx). - epics[].tasks[] — child tasks with the same fields; their
depends_onmay reference sibling tasks or other epics. - See the header comment in
roadmap.yamlfor the full field reference.
The cross-epic dependency graph — one node per epic, edges point prerequisite → dependent. Task-level detail lives in the collapsible sections below, and dependency-free epics are listed under the graph rather than drawn as disconnected boxes, so this stays legible at default zoom. Node colour = status: 🟢 done · 🔵 in-progress · 🟡 in-review · 🔴 blocked · ⚪ todo · ⚫ parked.
graph LR
sandbox_isolation["Non-Docker sandbox isolation (Landlock)"]
scanner_v2["Spec 076 deterministic offline tool-sca…"]
ux_audit["Web UI + macOS app UX audit"]
action_log_transparency["Action log / transparency — info at a g…"]
analytics_dashboard["Analytics dashboard as default page"]
scanner_simplification["Scanner simplification (deterministic d…"]
tpa_db["tpa-db: versioned TPA signature databas…"]
remote_access_tunnel["Remote access tunnel (feature-flagged M…"]
telemetry_identity["Telemetry identity & data quality (mach…"]
telemetry_v7_churn["Telemetry v7: honest funnel + churn ins…"]
scanner_v2 --> sandbox_isolation
ux_audit --> action_log_transparency
ux_audit --> analytics_dashboard
scanner_v2 --> scanner_simplification
scanner_simplification --> tpa_db
tpa_db --> remote_access_tunnel
ux_audit --> remote_access_tunnel
analytics_dashboard --> remote_access_tunnel
telemetry_identity --> telemetry_v7_churn
classDef done fill:#1f7a1f,stroke:#0d3d0d,color:#ffffff;
classDef in_progress fill:#1f6feb,stroke:#0b3d91,color:#ffffff;
classDef todo fill:#6e7781,stroke:#3d4248,color:#ffffff;
class sandbox_isolation,scanner_v2,scanner_simplification done;
class analytics_dashboard,telemetry_identity,telemetry_v7_churn in_progress;
class ux_audit,action_log_transparency,tpa_db,remote_access_tunnel todo;
Independent epics (15) — no cross-epic prerequisites; each stands alone:
- 🔵 Release qualification gate (auto-QA matrix blocks the tag) — In progress · P0
- 🟡 Windows native tray app — In review · P2
- 🔴 MCP protocol upgrade to 2026-07-28 revision — Blocked · P3
- ⚪ Planning/docs truth automation — Todo · P2
- ⚫ Server marketplace — Todo · P3 · parked
- ⚫ Audit SIEM integration — Todo · P3 · parked
- ⚫ Paid-tier MVP (billing / seats / license) — Todo · P3 · parked
- ⚫ SDK v1 migration — Todo · P3 · parked
- ⚫ SSO (server edition) — Todo · P3 · parked
- ⚪ Security gateway Tracks C/D (per-arg least-privilege + signature provenance) — Todo · P3
- ⚪ Discovery-quality eval harness (Spec 065 second half) — Todo · P3
- 🟢 Upgrade awareness & guided update — Done · P0
- 🟢 Connect step trust: preview, visible backup, one-click undo — Done · P0
- 🟢 Registries — easier search + add-server — Done · P1
- 🟢 Tray↔core decoupling: socket/REST API only, no config-file reads — Done · P2
Each epic's child tasks, their internal dependency graph, and tracker/PR links — collapsed by default, expand the ones you care about. Full metadata (priority, spec progress) is in the Epics table below.
🔵 Release qualification gate (auto-QA matrix blocks the tag) — In progress · P0
Attacks the return cliff (48% of installs are one-and-done; day-1 return 31% — corrected 2026-07-10, identity-deduped; the earlier '17.7% day-2' figure was un-deduped anonymous_id churn) and the one conceded competitor advantage: stability. No release tag until the surface x server-type matrix (MCP/REST/CLI/Web UI x stdio/http/sse/docker/oauth) plus invariants (activity-log/token counters move, quarantine flow, reconnect survival, in-place upgrade) pass automatically; macOS app smoke is advisory until promoted (3 consecutive passes, spec 081 US4). Assembles existing assets: test-api-e2e.sh, Playwright sweep, scan-eval gate, mcpproxy-ui-test.
Spec: 081-release-qa-gate
graph LR
release_qa_gate_matrix["T1: tag-blocking release-gate workflow: serve…"]
release_qa_gate_playwright["T2: wire the Playwright Web UI sweep into the…"]
release_qa_gate_macos["T3: macOS app smoke on a macos runner, adviso…"]
release_qa_gate_consistency["T4: surface-state consistency check (tray/Web…"]
release_qa_gate_matrix --> release_qa_gate_playwright
release_qa_gate_matrix --> release_qa_gate_macos
release_qa_gate_matrix --> release_qa_gate_consistency
classDef done fill:#1f7a1f,stroke:#0d3d0d,color:#ffffff;
classDef todo fill:#6e7781,stroke:#3d4248,color:#ffffff;
class release_qa_gate_matrix done;
class release_qa_gate_playwright,release_qa_gate_macos,release_qa_gate_consistency todo;
🔵 Analytics dashboard as default page — In progress · P1
Per-server / per-tool token-drain graphs; make the dashboard the default landing page. 2026-07-10 truth-sync: spec 069 is SHIPPED (25/26 — the only open task is a Playwright verification sweep), so the graphs half is done; only the default-landing half remains.
Spec: 069-observability-usage-graphs
graph LR
analytics_token_drain_graphs["Per-server / per-tool token-drain graphs"]
analytics_default_landing["Make dashboard the default landing page"]
analytics_token_drain_graphs --> analytics_default_landing
classDef done fill:#1f7a1f,stroke:#0d3d0d,color:#ffffff;
classDef todo fill:#6e7781,stroke:#3d4248,color:#ffffff;
class analytics_token_drain_graphs done;
class analytics_default_landing todo;
| Task | Status | Refs |
|---|---|---|
| Per-server / per-tool token-drain graphs | 🟢 Done | — |
| Make dashboard the default landing page | ⚪ Todo | — |
🔵 Telemetry identity & data quality (machine_id + CI-filter hardening) — In progress · P1
2026-07-11 source audit: the CLIENT half is DONE and RELEASED — the old framing ('add a hashed machine_id (schema v6)') is stale. machine_id (HMAC-SHA256 of the OS machine id, internal/telemetry/machine_id.go) is emitted unconditionally in every heartbeat (telemetry.go:789; telemetry is opt-out) and shipped in v0.47.0; the client schema is already v7, a version past the v6 this note described; CI is filtered client-side by disabling telemetry outright (env_overrides.go). Worker verified live in prod 2026-07-07 (machine_id 100% populated). The audit also found that the 79%-unknown launch_source was a CLIENT bug misfiled to mcpproxy-dash — a dashboard cannot display a value the client is incapable of sending. That is now fixed (see telemetry-launchsource-tray). Remaining: dashboard consumption (mcpproxy-dash) + snapshot-cron alerting.
graph LR
telemetry_machineid_client["Hashed machine_id in heartbeat (schema v6)"]
telemetry_machineid_worker["Worker migration: machine_id column + extract…"]
telemetry_launchsource_tray["Emit launch_source=tray — the 'tray' value wa…"]
telemetry_machineid_dash["Dashboard identityExpr prefers machine_id; ex…"]
telemetry_snapshot_alerting["Alerting on external-downloads snapshot cron…"]
telemetry_machineid_client --> telemetry_machineid_worker
telemetry_machineid_worker --> telemetry_machineid_dash
telemetry_launchsource_tray --> telemetry_machineid_dash
classDef done fill:#1f7a1f,stroke:#0d3d0d,color:#ffffff;
classDef todo fill:#6e7781,stroke:#3d4248,color:#ffffff;
class telemetry_machineid_client,telemetry_machineid_worker,telemetry_launchsource_tray done;
class telemetry_machineid_dash,telemetry_snapshot_alerting todo;
| Task | Status | Refs |
|---|---|---|
| Hashed machine_id in heartbeat (schema v6) | 🟢 Done | smart-mcp-proxy#796 |
| Worker migration: machine_id column + extraction (repo mcpproxy-telemetry) | 🟢 Done | mcpproxy-telemetry#3 |
| Emit launch_source=tray — the 'tray' value was UNREACHABLE in the client, and that (not the dashboard) was the 79%-unknown root cause | 🟢 Done | — |
| Dashboard identityExpr prefers machine_id; exclude %-dev versions from human cohort (repo mcpproxy-dash) | ⚪ Todo | — |
| Alerting on external-downloads snapshot cron (34-day outage went unnoticed) | ⚪ Todo | — |
🔵 Telemetry v7: honest funnel + churn instrumentation — In progress · P1
2026-07-06 recheck DEBUNKED the 72.4% connect-skip story: genuine never-connected skip = 0% (wizard dismiss stamped 'skipped' on users who connected via ConnectModal/CLI/manual config). 2026-07-10 recheck debunked the OTHER two spec-080 headline metrics as well: 'day-2 return 17.7%' was un-deduped anonymous_id churn (true, identity-deduped: one-and-done 48%, day-1 return 31%, day-7 16.6% — matches dashboard); '42% retrieve_tools -> 16% real call' was lifetime-flag vs windowed-counter asymmetry (true conversion ~90%; missing piece is a first_real_tool_call_ever activation flag). Real cliff = 48% one-and-done. 2026-07-11 source audit + fix: ALL of spec 080 is shipped and released in v0.47.0 (#813) — T1 wizard completed_external, T2 funnel fields, T3 pre-churn snapshot, US4 schema v7. The previous note claimed 'only cross-repo churn analytics (T4) remains'; that was WRONG — first_real_tool_call_ever had ZERO occurrences in Go code and was in-repo CLIENT work. It is now implemented, so the retrieve->call funnel is finally measurable lifetime-vs-lifetime. P0->P1: only cross-repo T4 now remains.
Spec: 080-telemetry-v7-churn
graph LR
telemetry_v7_wizard_fix["T1: wizard metric fix - on dismiss record con…"]
telemetry_v7_funnel_fields["T2: funnel observability fields (wizard_shown…"]
telemetry_v7_prechurn_snapshot["T3: pre-churn snapshot (previous_shutdown cle…"]
telemetry_v7_realcall_flag["first_real_tool_call_ever activation flag (sy…"]
telemetry_v7_churn_events["T4: cross-repo churn_events materialization +…"]
telemetry_v7_wizard_fix --> telemetry_v7_churn_events
telemetry_v7_funnel_fields --> telemetry_v7_churn_events
telemetry_v7_prechurn_snapshot --> telemetry_v7_churn_events
classDef done fill:#1f7a1f,stroke:#0d3d0d,color:#ffffff;
classDef todo fill:#6e7781,stroke:#3d4248,color:#ffffff;
class telemetry_v7_wizard_fix,telemetry_v7_funnel_fields,telemetry_v7_prechurn_snapshot,telemetry_v7_realcall_flag done;
class telemetry_v7_churn_events todo;
| Task | Status | Refs |
|---|---|---|
| T1: wizard metric fix - on dismiss record connect step as completed_external (not skipped) when the user already connected via another path | 🟢 Done | #813 |
| T2: funnel observability fields (wizard_shown, web_ui_opened counter, days_since_install, active_days_30d) | 🟢 Done | #813 |
| T3: pre-churn snapshot (previous_shutdown clean|crash via BBolt flag, last_error_code) so the final heartbeat doubles as cause-of-death | 🟢 Done | #813 |
| first_real_tool_call_ever activation flag (symmetric to first_retrieve_tools_call_ever) so the retrieve->call funnel step is measurable lifetime-vs-lifetime | 🟢 Done | — |
| T4: cross-repo churn_events materialization + dash Churn page with H1-H4 hypothesis signatures (repos mcpproxy-telemetry / mcpproxy-dash; tracked here for DAG visibility, out of scope of spec 080) | ⚪ Todo | — |
🟡 Windows native tray app — In review · P2 · MCP-43
No spec: link — this epic is the native TRAY app; specs/002-windows-installer is the unrelated INSTALLER spec (35/60) and its badge said nothing about tray progress (wrong link removed 2026-07-10). Option C: WebView2 window reusing shipped Web UI. Most exit criteria already ship; gaps = native window, toasts, profile submenu, Win11 smoke. Telemetry: Windows = ~23% of GitHub downloads but only ~4% of active installs (downloads→actives ~12:1 vs macOS ~4:1) — gate WebView2 work on finding the funnel break first.
graph LR
windows_tray_funnel_qa["Windows first-run QA pass (downloads→actives…"]
windows_tray_window["WebView2 native window + profile submenu<br/>MCP-43"]
windows_tray_funnel_qa --> windows_tray_window
classDef in_review fill:#9a6700,stroke:#5c3d00,color:#ffffff;
classDef todo fill:#6e7781,stroke:#3d4248,color:#ffffff;
class windows_tray_window in_review;
class windows_tray_funnel_qa todo;
| Task | Status | Refs |
|---|---|---|
| Windows first-run QA pass (downloads→actives 12:1 vs macOS 4:1 — find the funnel break before WebView2 work) | ⚪ Todo | — |
| WebView2 native window + profile submenu | 🟡 In review | MCP-43 |
🔴 MCP protocol upgrade to 2026-07-28 revision — Blocked · P3
BLOCKED on mcp-go shipping 2026-07-28 (pinned v0.55.x tops out at 2025-11-25). CROSS-SPEC CONFLICT: FR-012 forbids per-connection */list variation; SHIPPED Spec 057 selects toolset by URL path /mcp/p/. Must reconcile at plan time (058 spec now carries a Cross-Spec Reconciliation note). 028 agent-token scoping is already compatible (header-carried).
Spec: 058-mcp-2026-upgrade
⚪ Web UI + macOS app UX audit — Todo · P0
End-to-end UX pass across Web UI and the macOS tray app; the umbrella for the polish push. (No spec yet — 064 is the unrelated agent-fleet glass-cockpit spec.)
graph LR
ux_audit_webui_sweep["Web UI heuristic + Playwright UX sweep"]
ux_audit_macos_sweep["macOS tray app UX sweep (settings parity, flo…"]
classDef todo fill:#6e7781,stroke:#3d4248,color:#ffffff;
class ux_audit_webui_sweep,ux_audit_macos_sweep todo;
| Task | Status | Refs |
|---|---|---|
| Web UI heuristic + Playwright UX sweep | ⚪ Todo | — |
| macOS tray app UX sweep (settings parity, flows) | ⚪ Todo | — |
⚪ Action log / transparency — info at a glance — Todo · P1
Surface the most important activity/security/connection signals at a glance; reduce digging. Vision pillar 'feel control → transparency' — the activity log is a headline feature, polish it and bring it to the tray menu. Builds on the shipped activity-log backend + retention (spec 024, 95% shipped — this epic is the at-a-glance UX on top, not the backend, so 024 is not the progress driver).
graph LR
sessions_web_ui["Sessions in the Web UI: meaningful session na…"]
action_log_glance_view["At-a-glance action log view (top signals, hea…"]
action_log_tray_menu["Activity in the tray menu (recent tool calls…"]
tray_menu_open_telemetry["tray_menu_opened counter: Swift menuWillOpen…"]
action_log_retention_tie_in["Tie activity retention/size into the glance v…"]
action_log_glance_view --> action_log_tray_menu
action_log_glance_view --> action_log_retention_tie_in
classDef done fill:#1f7a1f,stroke:#0d3d0d,color:#ffffff;
classDef todo fill:#6e7781,stroke:#3d4248,color:#ffffff;
class sessions_web_ui done;
class action_log_glance_view,action_log_tray_menu,tray_menu_open_telemetry,action_log_retention_tie_in todo;
| Task | Status | Refs |
|---|---|---|
| Sessions in the Web UI: meaningful session names in the Activity Log filter + the existing /sessions page linked in the sidebar | 🟢 Done | — |
| At-a-glance action log view (top signals, health) | ⚪ Todo | — |
| Activity in the tray menu (recent tool calls + security events, jump to full log) | ⚪ Todo | — |
| tray_menu_opened counter: Swift menuWillOpen (MCPProxyApp.swift:192) -> lightweight POST /api/v1/telemetry/tray-menu-opened -> registry counter -> heartbeat tray_menu_opened_24h | ⚪ Todo | — |
| Tie activity retention/size into the glance view | ⚪ Todo | — |
⚪ tpa-db: versioned TPA signature database for the offline scanner — Todo · P1
Vision pillar 'feel protected': the deterministic detect engine (Spec 076/077) ships with built-in checks but no updatable knowledge of in-the-wild Tool Poisoning Attacks. Build a versioned, offline-first signature/pattern database (known TPA campaigns, malicious phrase corpora, IoC hashes) that the engine consumes — bundled with the binary, refreshable out-of-band, community-contributable, and guarded by the existing scan-eval recall/FP CI gate.
graph LR
tpa_db_format["Signature DB format + loader (versioned, sign…"]
tpa_db_corpus["Seed corpus: catalog known public TPA campaig…"]
tpa_db_refresh["Out-of-band refresh (offline-friendly: manual…"]
tpa_db_format --> tpa_db_corpus
tpa_db_format --> tpa_db_refresh
classDef todo fill:#6e7781,stroke:#3d4248,color:#ffffff;
class tpa_db_format,tpa_db_corpus,tpa_db_refresh todo;
| Task | Status | Refs |
|---|---|---|
| Signature DB format + loader (versioned, signed, bundled default) | ⚪ Todo | — |
| Seed corpus: catalog known public TPA campaigns/patterns into the DB | ⚪ Todo | — |
| Out-of-band refresh (offline-friendly: manual file drop + optional fetch), eval-gated | ⚪ Todo | — |
⚪ Remote access tunnel (feature-flagged MVP, spec 089) — Todo · P2
One-button Web UI exposure of /mcp via external tunnel binary (cloudflared quick tunnel first) so Claude custom connectors (all tiers incl. Free, syncs to iOS/Android) can reach local MCP servers (e.g. Obsidian) — behind a feature flag, off by default, mandatory OAuth 2.1+PKCE+DCR gate, per-server exposure allowlist, remote-origin activity logging. Research: docs/research/remote-access-tunnel-research-2026-07-29.html (25/25 claims verified; niche unoccupied — Docker MCP Gateway lacks it). Sequenced after tpa-db (+ shipped scanner work 086-088), macOS tray redesign (ux-audit) and analytics-dashboard per owner decision 2026-07-29. No hosted relay/payments in MVP.
Spec: 089-remote-access-tunnel
graph LR
tunnel_oauth_gate["OAuth 2.1 authorization-server gate for tunne…"]
tunnel_orchestration["cloudflared quick-tunnel orchestration (detec…"]
tunnel_exposure_allowlist["Per-server exposure allowlist (default none;…"]
tunnel_webui_tray["Web UI open/close button + URL/QR/instruction…"]
tunnel_oauth_gate --> tunnel_exposure_allowlist
tunnel_orchestration --> tunnel_webui_tray
tunnel_exposure_allowlist --> tunnel_webui_tray
classDef todo fill:#6e7781,stroke:#3d4248,color:#ffffff;
class tunnel_oauth_gate,tunnel_orchestration,tunnel_exposure_allowlist,tunnel_webui_tray todo;
| Task | Status | Refs |
|---|---|---|
| OAuth 2.1 authorization-server gate for tunnel-origin traffic (PKCE, DCR, Anthropic callback allowlist, token lifecycle/revocation) | ⚪ Todo | — |
| cloudflared quick-tunnel orchestration (detect/launch/supervise/parse URL) + feature flag + never-auto-start | ⚪ Todo | — |
| Per-server exposure allowlist (default none; quarantined non-exposable; hot-reload) | ⚪ Todo | — |
| Web UI open/close button + URL/QR/instructions + warning banner; tray active-state indicator; remote-origin activity marker | ⚪ Todo | — |
⚪ Planning/docs truth automation — Todo · P2
Automate the consistency checks this very audit had to do by hand: roadmap vs GitHub PR state, tasks.md updates on implementation PRs, volatile CLAUDE.md/README facts, and quickstart contract tests.
graph LR
hygiene_roadmap_github_check["gen-roadmap --check-github: cross-check roadm…"]
hygiene_tasks_reconcile["CI rule: PR touching specs/<id> implementatio…"]
hygiene_spec_evidence_check["scripts/check-spec-evidence.py: deterministic…"]
hygiene_spec_gardener["Weekly cloud routine: LLM judges only the res…"]
hygiene_docs_facts["Generate volatile CLAUDE.md/README facts (Go…"]
hygiene_quickstart_contract["Run top quickstart.md scenario per spec as co…"]
hygiene_spec_evidence_check --> hygiene_spec_gardener
classDef done fill:#1f7a1f,stroke:#0d3d0d,color:#ffffff;
classDef in_progress fill:#1f6feb,stroke:#0b3d91,color:#ffffff;
classDef todo fill:#6e7781,stroke:#3d4248,color:#ffffff;
class hygiene_roadmap_github_check,hygiene_spec_evidence_check done;
class hygiene_spec_gardener in_progress;
class hygiene_tasks_reconcile,hygiene_docs_facts,hygiene_quickstart_contract todo;
| Task | Status | Refs |
|---|---|---|
| gen-roadmap --check-github: cross-check roadmap.yaml statuses vs gh PR state + dangling spec links | 🟢 Done | #800 |
| CI rule: PR touching specs/ implementation paths must update tasks.md | ⚪ Todo | — |
| scripts/check-spec-evidence.py: deterministic check that every TICKED task cites code that exists | 🟢 Done | — |
| Weekly cloud routine: LLM judges only the residue the evidence-check cannot decide, opens/updates one propose-only PR | 🔵 In progress | — |
| Generate volatile CLAUDE.md/README facts (Go version, built-in tool list, sample config) from code with --check | ⚪ Todo | — |
| Run top quickstart.md scenario per spec as contract test in test-api-e2e.sh | ⚪ Todo | — |
⚪ Security gateway Tracks C/D (per-arg least-privilege + signature provenance) — Todo · P3
Track A→Spec 056, Track B→Spec 059 (both shipped). UNBUILT: Track C per-ARGUMENT allow-listing (per-tool scope exists in mcp_direct_scope.go); Track D provenance + human-readable signature diff (SHA-256 pinning exists via Spec 032). Build ON 032/028, don't re-implement; honor the rug-pull re-quarantine interaction rule vs 032 auto-approve.
Spec: 054-mcp-security-gateway
⚪ Discovery-quality eval harness (Spec 065 second half) — Todo · P3
Security recall/FP half SHIPPED (cmd/scan-eval, backs Spec 076/077 gate). UNBUILT: the discovery-quality (retrieve_tools recall) eval harness.
⚫ Server marketplace — Todo · parked · P3 · MCP-37
PARKED. ~60% already ships (browse/search/one-click add). No spec yet; gaps tracked as MCP-3246..3250 (tray entries, metadata, telemetry). (070 is the registries-search-add spec, not a marketplace spec.)
⚫ Audit SIEM integration — Todo · parked · P3 · MCP-39
PARKED. Splunk HEC / Elastic _bulk / syslog shippers reusing JSONL export pipeline.
⚫ Paid-tier MVP (billing / seats / license) — Todo · parked · P3 · MCP-40
PARKED. Server-edition revenue motion: Ed25519 license tokens, seats, Stripe checkout. Behind //go:build server.
⚫ SDK v1 migration — Todo · parked · P3
PARKED. Migrate to the v1 MCP Go SDK surface.
⚫ SSO (server edition) — Todo · parked · P3
PARKED. Single sign-on for the multi-user server edition.
🟢 Upgrade awareness & guided update — Done · P0
Corrected CI-filtered telemetry (2026-07-02): ~60% of last-14d active installs run pre-v0.40; latest stable v0.46.0 only 18.7%. Turn the existing internal/updatecheck background poll into a universal, non-intrusive, channel-aware upgrade nudge across every surface. Never blocks/modals; silent offline/CI.
Spec: 079-upgrade-nudge
graph LR
upgrade_nudge_status_log["US1 slice: update availability in mcpproxy st…"]
upgrade_nudge_surfacing["US1 remainder: dismissible Web UI banner + up…"]
upgrade_nudge_channel["US2: channel-aware guided update command (bre…"]
upgrade_nudge_quiet["US3: operator control + CI/offline quiet + no…"]
upgrade_nudge_status_log --> upgrade_nudge_surfacing
upgrade_nudge_surfacing --> upgrade_nudge_channel
upgrade_nudge_surfacing --> upgrade_nudge_quiet
classDef done fill:#1f7a1f,stroke:#0d3d0d,color:#ffffff;
class upgrade_nudge_status_log,upgrade_nudge_surfacing,upgrade_nudge_channel,upgrade_nudge_quiet done;
| Task | Status | Refs |
|---|---|---|
| US1 slice: update availability in mcpproxy status + deduped startup log | 🟢 Done | #798 |
| US1 remainder: dismissible Web UI banner + update_check config block | 🟢 Done | #805 |
| US2: channel-aware guided update command (brew/dmg/deb/rpm/docker/go-install detection, build-time channel marker) | 🟢 Done | #818 |
| US3: operator control + CI/offline quiet + no prerelease downgrade nudges | 🟢 Done | #911 |
🟢 Connect step trust: preview, visible backup, one-click undo — Done · P0
Legacy wizard telemetry APPEARED to show 72.4% of engaged users skipping the connect step - debunked 2026-07-06: an instrumentation artifact, genuine never-connected skip = 0% (the wizard stamped skipped on users who connected via ConnectModal/CLI/manual config); real cliff is one-and-done installs ~48% (day-1 return 31%, identity-deduped 2026-07-10), see specs/080. Completers retain ~50% at two weeks vs 6% for non-engaged (correlation with engagement, not causation by the connect step). Backups already exist (internal/connect/backup.go) but are invisible in the Web UI. Close the trust gap: preview the exact config diff, surface the backup, offer one-click undo, explain the macOS TCC prompt.
Spec: 078-connect-trust-preview
graph LR
connect_trust_preview["US1: preview API + wizard diff UI (exact entr…"]
connect_trust_backup_visibility["US1: surface backup_path in Web UI + retentio…"]
connect_trust_undo["US2: one-click undo/disconnect in wizard"]
connect_trust_tcc_copy["US2: pre-emptive macOS TCC explanation in wiz…"]
classDef done fill:#1f7a1f,stroke:#0d3d0d,color:#ffffff;
class connect_trust_preview,connect_trust_backup_visibility,connect_trust_undo,connect_trust_tcc_copy done;
| Task | Status | Refs |
|---|---|---|
| US1: preview API + wizard diff UI (exact entry, API-key masking) | 🟢 Done | #802 |
| US1: surface backup_path in Web UI + retention policy | 🟢 Done | #799 |
| US2: one-click undo/disconnect in wizard | 🟢 Done | #804 |
| US2: pre-emptive macOS TCC explanation in wizard | 🟢 Done | #910 |
🟢 Non-Docker sandbox isolation (Landlock) — Done · P1 · MCP-34
Landlock LSM + setrlimit native sandbox for stdio upstreams; no userns (Ubuntu 24.04 safe). Originated from roadmap item #11 (no dedicated spec — 054 is the unrelated security-gateway spec). Code in internal/sandbox/; PRs #754/#759/#768/#781/#782.
graph LR
sandbox_spike["Landlock sandbox spike (MCP-34.1)<br/>MCP-3232"]
sandbox_mode_config["isolation.mode enum + resolver (MCP-34.2)<br/>MCP-3233"]
sandbox_launcher["Native sandbox launcher Landlock+rlimits (MCP…<br/>MCP-3234"]
sandbox_scanner_parity["Scanner-flow parity under sandbox (MCP-34.4)<br/>MCP-3235"]
sandbox_snap_docker_it["snap-docker integration tests + CI (MCP-34.5)<br/>MCP-3236"]
sandbox_spike --> sandbox_mode_config
sandbox_mode_config --> sandbox_launcher
sandbox_launcher --> sandbox_scanner_parity
sandbox_scanner_parity --> sandbox_snap_docker_it
classDef done fill:#1f7a1f,stroke:#0d3d0d,color:#ffffff;
class sandbox_spike,sandbox_mode_config,sandbox_launcher,sandbox_scanner_parity,sandbox_snap_docker_it done;
| Task | Status | Refs |
|---|---|---|
| Landlock sandbox spike (MCP-34.1) | 🟢 Done | MCP-3232 #754 |
| isolation.mode enum + resolver (MCP-34.2) | 🟢 Done | MCP-3233 #759 |
| Native sandbox launcher Landlock+rlimits (MCP-34.3) | 🟢 Done | MCP-3234 #768 |
| Scanner-flow parity under sandbox (MCP-34.4) | 🟢 Done | MCP-3235 #781 |
| snap-docker integration tests + CI (MCP-34.5) | 🟢 Done | MCP-3236 #782 |
🟢 Spec 076 deterministic offline tool-scanner — Done · P1 · MCP-3574
Deterministic offline signal pipeline replaces ~10%-recall scanner; scan-eval --gate (recall>=0.90 / FP<=5%) in CI.
Spec: 076-deterministic-tool-scanner
graph LR
scanner_v2_foundation["detect-engine foundation (T1)<br/>MCP-3575"]
scanner_v2_hard_checks["3 hard checks + scanner wiring (US1 MVP)<br/>MCP-3576"]
scanner_v2_soft_checks["3 soft checks + patterns confidence (US2)<br/>MCP-3577"]
scanner_v2_consensus["Consensus risk-score + report transparency (U…<br/>MCP-3578"]
scanner_v2_eval_gate["Eval corpus + CI recall/FP gate (US3)<br/>MCP-3579"]
scanner_v2_docs["Tool-scanner detect-engine docs (T22)<br/>MCP-3683"]
scanner_v2_foundation --> scanner_v2_hard_checks
scanner_v2_foundation --> scanner_v2_soft_checks
scanner_v2_hard_checks --> scanner_v2_consensus
scanner_v2_soft_checks --> scanner_v2_consensus
scanner_v2_hard_checks --> scanner_v2_eval_gate
scanner_v2_eval_gate --> scanner_v2_docs
classDef done fill:#1f7a1f,stroke:#0d3d0d,color:#ffffff;
class scanner_v2_foundation,scanner_v2_hard_checks,scanner_v2_soft_checks,scanner_v2_consensus,scanner_v2_eval_gate,scanner_v2_docs done;
| Task | Status | Refs |
|---|---|---|
| detect-engine foundation (T1) | 🟢 Done | MCP-3575 #769 |
| 3 hard checks + scanner wiring (US1 MVP) | 🟢 Done | MCP-3576 #770 |
| 3 soft checks + patterns confidence (US2) | 🟢 Done | MCP-3577 #775 |
| Consensus risk-score + report transparency (US4) | 🟢 Done | MCP-3578 #776 |
| Eval corpus + CI recall/FP gate (US3) | 🟢 Done | MCP-3579 #777 |
| Tool-scanner detect-engine docs (T22) | 🟢 Done | MCP-3683 #780 |
🟢 Registries — easier search + add-server — Done · P1
Lower the friction of finding a server in a registry and adding it; lean on the official registry protocol work. 2026-07-10 truth-sync: both children shipped — spec 070 is 21/24 (the 3 open tasks are pre-PR chores: worktree baseline, run gates, apply gate decisions) and 071 is 12/12. depends_on [ux-audit] dropped: a done epic cannot depend on a todo one.
Spec: 070-registry-easy-upstream-add
graph LR
registries_search_ux["Improved registry search UX"]
registries_official_protocol["Official registry protocol integration"]
classDef done fill:#1f7a1f,stroke:#0d3d0d,color:#ffffff;
class registries_search_ux,registries_official_protocol done;
| Task | Status | Refs |
|---|---|---|
| Improved registry search UX | 🟢 Done | — |
| Official registry protocol integration | 🟢 Done | — |
🟢 Scanner simplification (deterministic default, opt-in deep scan) — Done · P1
Make the Spec 076 detect engine the always-on offline default; demote Docker scanners + source extraction to opt-in deep scan that never blocks/degrades the baseline; single unified report. COMPLETE: US1 #786, US2 #792, US4 #794, US3 + deep-scan trust fixes + docs truth sweep (T037-T039) #793 — all merged; shipped in v0.47.0-rc.2. Remaining 4 unchecked tasks in tasks.md are documented scope-outs. First of the 5 personal-edition polish verticals.
Spec: 077-scanner-simplification
graph LR
scanner_simpl_baseline["US1: deterministic offline baseline default +…"]
scanner_simpl_unified_report["US2: single merged report + cross-scanner con…"]
scanner_simpl_deep_optin["US3: opt-in deep scan (off by default), never…"]
scanner_simpl_notifications["US4: collapse scan-notification storm into on…"]
scanner_simpl_deepscan_fixes["Deep-scan trust fixes: nil-Security gating bu…"]
scanner_simpl_baseline --> scanner_simpl_unified_report
scanner_simpl_baseline --> scanner_simpl_deep_optin
scanner_simpl_unified_report --> scanner_simpl_deep_optin
scanner_simpl_unified_report --> scanner_simpl_notifications
scanner_simpl_deep_optin --> scanner_simpl_deepscan_fixes
classDef done fill:#1f7a1f,stroke:#0d3d0d,color:#ffffff;
class scanner_simpl_baseline,scanner_simpl_unified_report,scanner_simpl_deep_optin,scanner_simpl_notifications,scanner_simpl_deepscan_fixes done;
| Task | Status | Refs |
|---|---|---|
| US1: deterministic offline baseline default + curated hard phrase_injection check (delete duplicate legacy rules) | 🟢 Done | #786 |
| US2: single merged report + cross-scanner consensus confidence | 🟢 Done | #792 |
| US3: opt-in deep scan (off by default), never blocks/degrades baseline; config migration | 🟢 Done | #793 |
| US4: collapse scan-notification storm into one debounced settled event (MCP-2207) | 🟢 Done | #794 |
| Deep-scan trust fixes: nil-Security gating bug (source fetch runs with deep scan off on default configs), FR-014 verdict inversion (Dangerous deep finding < Warning), surface silently-skipped Docker scanners (non-nil deep_scan descriptor + CLI hint on security enable) | 🟢 Done | #793 |
🟢 Tray↔core decoupling: socket/REST API only, no config-file reads — Done · P2
Architecture rule (CLAUDE.md): the tray holds no state and talks to the core only via socket/REST + SSE. 2026-07-11 source-of-truth re-audit + fix: Swift tray was already clean (MCPProxyApp.swift opens the config in an external editor, never parses it); the Go tray's update-check gate was already reworked to core-API gating (#805). The last violation — config.LoadFromFile in the Go tray's OAuth login path, live since ff03db92 (2026-05-18, #477) — turned out to be FUNCTIONALLY DEAD: the loaded config fed only two debug log lines, while the actual trigger was already the core-API TriggerOAuthLogin. Deleted rather than ported to REST. Bootstrap reads (socket path, config PATH without parsing, CA cert) are allowed and remain. Now enforced by a test so the rule cannot silently rot.
graph LR
tray_oauth_config_read["Delete the dead config read in the Go tray OA…"]
tray_config_import_guard["Test guard (internal/tray/config_import_guard…"]
tray_oauth_config_read --> tray_config_import_guard
classDef done fill:#1f7a1f,stroke:#0d3d0d,color:#ffffff;
class tray_oauth_config_read,tray_config_import_guard done;
| Task | Status | Refs |
|---|---|---|
| Delete the dead config read in the Go tray OAuth path (config.LoadFromFile) + drop the now-unused internal/config import. GetConfigPath stays on the interface — openConfigDir still needs the path to reveal the dir in the file manager. | 🟢 Done | — |
| Test guard (internal/tray/config_import_guard_test.go) failing any tray-side call to config.LoadFromFile/Load/SaveConfig/... Parses source on disk, so a violation cannot hide behind an inactive build tag. Bans config FILE I/O, not the config package — cmd/mcpproxy-tray legitimately references the config.LogConfig type for its own logger. | 🟢 Done | — |
| Epic | Status | Priority | Progress | Spec | PR |
|---|---|---|---|---|---|
| Release qualification gate (auto-QA matrix blocks the tag) | In progress | P0 | — | 081-release-qa-gate | |
| Analytics dashboard as default page | In progress | P1 | 25/26 (96%) | 069-observability-usage-graphs | |
| Telemetry identity & data quality (machine_id + CI-filter hardening) | In progress | P1 | — | ||
| Telemetry v7: honest funnel + churn instrumentation | In progress | P1 | — | 080-telemetry-v7-churn | |
Windows native tray app MCP-43 |
In review | P2 | — | ||
| MCP protocol upgrade to 2026-07-28 revision | Blocked | P3 | — | 058-mcp-2026-upgrade | |
| Web UI + macOS app UX audit | Todo | P0 | — | ||
| Action log / transparency — info at a glance | Todo | P1 | — | ||
| tpa-db: versioned TPA signature database for the offline scanner | Todo | P1 | — | ||
| Remote access tunnel (feature-flagged MVP, spec 089) | Todo | P2 | — | 089-remote-access-tunnel | |
| Planning/docs truth automation | Todo | P2 | — | ||
| Security gateway Tracks C/D (per-arg least-privilege + signature provenance) | Todo | P3 | — | 054-mcp-security-gateway | |
| Discovery-quality eval harness (Spec 065 second half) | Todo | P3 | — | 065-evaluation-foundation | |
Server marketplace MCP-37 |
Todo (parked) | P3 | — | ||
Audit SIEM integration MCP-39 |
Todo (parked) | P3 | — | ||
Paid-tier MVP (billing / seats / license) MCP-40 |
Todo (parked) | P3 | — | ||
| SDK v1 migration | Todo (parked) | P3 | — | ||
| SSO (server edition) | Todo (parked) | P3 | — | ||
| Upgrade awareness & guided update | Done | P0 | — | 079-upgrade-nudge | |
| Connect step trust: preview, visible backup, one-click undo | Done | P0 | — | 078-connect-trust-preview | |
Non-Docker sandbox isolation (Landlock) MCP-34 |
Done | P1 | — | ||
Spec 076 deterministic offline tool-scanner MCP-3574 |
Done | P1 | 22/24 (92%) | 076-deterministic-tool-scanner | |
| Registries — easier search + add-server | Done | P1 | 21/24 (88%) | 070-registry-easy-upstream-add | |
| Scanner simplification (deterministic default, opt-in deep scan) | Done | P1 | 38/42 (90%) | 077-scanner-simplification | |
| Tray↔core decoupling: socket/REST API only, no config-file reads | Done | P2 | — |
Swept out of the working set by --archive once done, merged and cooled off. Full entries — notes, child tasks, PR refs — live in roadmap.archive.yaml.
| Epic | Shipped | Archived | PRs |
|---|---|---|---|
Profiles v2 (per-profile tool views) MCP-33 |
2026-06-24 | 2026-07-10 | #756 #761 #766 #767 |
TypeScript code-execution GA + cookbook MCP-38 |
2026-06-24 | 2026-07-10 | #753 |
Legend: shipped ≥95% checked · in-flight 1–94% · drafted 0% · — no tasks.md. This aggregate is regenerated here rather than overwriting the hand-maintained specs/README.md, which keeps its curated prose, runbooks and design-doc links.