feat(profile): /api/profile + eligibility filter + inference framework (ADR-0014 steps 4-6)

Step 4 — /api/profile read-through API:
  GET  /api/profile          → { user, prefs, consents, contexts }
  PATCH /api/profile/prefs/:scope  upsert user_preferences (source='user')
  PATCH /api/profile/consents      grant / revoke consent keys
  PATCH /api/profile/contexts      create / activate / deactivate contexts
  Legacy consentGiven bit folded in as data:core fallback.

Step 5 — registry-driven eligibility filter:
  fetchRegistry() exported from agent-registry.ts.
  profile/eligibility.ts: getEligibleAgentIds(userId) — filters by required
  consents, silenced_in_contexts, and user_preferences[enabled=false].
  fetchOrchestratorTip filters agent_outputs to eligible set before calling
  ml/serving /recommend. Fail-closed: registry unavailable → empty set.

Step 6 — shared context-inference framework (#111) + time-of-day proof (#112):
  ml/agents/inference/: UserHistory, FeedbackEvent, run_inference().
  Framework: cold-start, min_history gating, error fallback, structured logs.
  TimeOfDayAgent v1.1.0: inferred_params=[preferred_hour]; also reads
  quiet_start/quiet_end from agent_prefs. agent_prefs injected by TS caller.
  AgentInput gains agent_prefs field.
  ml/serving: POST /agents/{agent_id}/infer endpoint.
  agent-outputs.ts computeAndStore: loads prefs before compute, calls /infer
  after, persists results (source='inferred'); user overrides never touched.

Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>

CLAUDE.md

@@ -106,15 +106,35 @@ Recent completions:
 - Model benchmarking for tip generation (#93, #95)
 - Admin UX refinements: feedback consolidation, settings placement (#100–102)
 - ADR-0012 — ε-greedy v2 (D=12) — 2026-04-26 (now superseded by ADR-0013)
+- ADR-0014 steps 1–6: unified Profile schema + backfill, manifest plumbing, `/api/profile` read-through, registry-driven eligibility filter, inference framework + time-of-day migration — 2026-05-05
 
 Active work (M2):
-- ADR-0014 (proposed) — unified Profile model + agent registry + inference framework
+- ADR-0014 step 7 — per-agent inference: focus-area (#113), momentum (#114), overdue-task (#115), recent-patterns (#116)
-- Unified Profile model: prefs, contexts, consents, registry plumbing, orchestrator cutover (#30)
+- ADR-0014 step 8 — drop `users.consentGiven` column
-- Shared context-inference framework for agents (#111)
-- Per-agent auto-inference: time-of-day (#112), focus-area (#113), momentum (#114), overdue-task (#115), recent-patterns (#116)
 - Signal abstraction for multi-source support (#78)
 - Per-user feature freshness SLAs (#61, ADR-0011 phase B)
 
+## ADR-0014 endpoint map (as of step 6)
+
+| Endpoint | Purpose |
+|----------|---------|
+| `GET /api/profile` | Read-through: user globals + prefs (by scope) + consents + contexts |
+| `PATCH /api/profile/prefs/:scope` | Upsert user_preferences rows (source='user') |
+| `PATCH /api/profile/consents` | Grant / revoke consent keys |
+| `PATCH /api/profile/contexts` | Create / activate / deactivate named contexts |
+| `GET /api/agents/registry` | Manifest list (proxy to ml/serving; 60 s cache) |
+| `POST /api/agents/:agentId/compute` | Internal: run agent compute for (user, agent) |
+| `POST /agents/{agent_id}/infer` *(ml/serving)* | Run inference framework → `{inferred_prefs}` |
+
+## Inference framework (ADR-0014 §3)
+
+Lives in `ml/agents/inference/`. `run_inference(manifest, history)` evaluates all `InferredParam` entries in the manifest and returns `{key: value}`. Rules:
+- Below `min_history` → emit `cold_start_default`
+- `infer()` error → emit `cold_start_default` (never crashes)
+- Results written to `user_preferences` with `source='inferred'`; keys with `source='user'` are never overwritten
+
+Time-of-day agent (`1.1.0`) is the proof agent (#112): infers `preferred_hour` (mode done-hour) and reads `quiet_start`/`quiet_end` from prefs.
+
 ## What NOT to do
 
 - Don't copy Todoist's data into our DB. Store the OAuth token + computed features/derivatives we need, fetch raw on demand.

ml/agents/base.py

@@ -15,6 +15,11 @@ class AgentInput:
     profile: dict[str, float | None]           # profile feature values keyed by feature name
     feedback_history: list[dict] = field(default_factory=list)  # [{action, dwell_ms, created_at}, …]
     now: datetime = field(default_factory=lambda: datetime.now(timezone.utc))
+    # Per-agent inferred/user prefs loaded from user_preferences (ADR-0014 §3).
+    # Keys match the agent's pref_schema + inferred_params. 'user' source takes
+    # precedence over 'inferred' source; the caller resolves priority before
+    # passing this dict in.
+    agent_prefs: dict = field(default_factory=dict)
 
 
 @dataclass

ml/agents/inference/__init__.py

@@ -0,0 +1,9 @@
+"""Shared context-inference framework (ADR-0014 §3, issue #111).
+
+Each agent's manifest declares InferredParams; this package owns the
+scheduling contract, history data model, and write path to user_preferences.
+"""
+from .framework import run_inference
+from .history import FeedbackEvent, UserHistory
+
+__all__ = ["run_inference", "FeedbackEvent", "UserHistory"]

ml/agents/inference/framework.py

@@ -0,0 +1,59 @@
+"""run_inference — core of the context-inference framework (ADR-0014 §3).
+
+Contract:
+    run_inference(manifest, history) → dict[key, value]
+
+Semantics:
+  - For each InferredParam in manifest.inferred_params:
+      - If len(history.events) < param.min_history → emit cold_start_default.
+      - Otherwise → call param.infer(history) and emit the result.
+  - Returns {key: value} ready for the caller to persist to user_preferences
+    with source='inferred'.
+  - User overrides (source='user') are handled by the caller's upsert logic;
+    this function has no DB access.
+"""
+from __future__ import annotations
+
+import logging
+import time
+from typing import Any
+
+from ..manifest import AgentManifest
+from .history import UserHistory
+
+log = logging.getLogger(__name__)
+
+
+def run_inference(manifest: AgentManifest, history: UserHistory) -> dict[str, Any]:
+    """Evaluate all InferredParams for an agent and return {key: inferred_value}."""
+    result: dict[str, Any] = {}
+    n = len(history.events)
+
+    for param in manifest.inferred_params:
+        t0 = time.monotonic()
+        if param.infer is None:
+            result[param.key] = param.cold_start_default
+            continue
+        if n < param.min_history:
+            value = param.cold_start_default
+            source = "cold_start"
+        else:
+            try:
+                value = param.infer(history)
+                source = "inferred"
+            except Exception as exc:
+                log.warning(
+                    "inference_error agent=%s param=%s error=%s — using cold_start_default",
+                    manifest.id, param.key, exc,
+                )
+                value = param.cold_start_default
+                source = "error_fallback"
+
+        latency_ms = round((time.monotonic() - t0) * 1000, 1)
+        log.info(
+            "inference_param agent=%s param=%s source=%s value=%r history_len=%d latency_ms=%s",
+            manifest.id, param.key, source, value, n, latency_ms,
+        )
+        result[param.key] = value
+
+    return result

ml/agents/inference/history.py

@@ -0,0 +1,29 @@
+"""UserHistory — normalised view of a user's feedback events for inference."""
+from __future__ import annotations
+
+from dataclasses import dataclass, field
+from datetime import datetime, timezone
+
+
+@dataclass
+class FeedbackEvent:
+    action: str              # 'done' | 'dismiss' | 'snooze' | 'helpful' | 'not_helpful'
+    dwell_ms: int | None
+    created_at: str          # ISO 8601
+
+    @property
+    def hour(self) -> int:
+        """Hour of day (0-23) when the feedback was recorded."""
+        try:
+            dt = datetime.fromisoformat(self.created_at.replace("Z", "+00:00"))
+        except ValueError:
+            return 12
+        if dt.tzinfo is None:
+            dt = dt.replace(tzinfo=timezone.utc)
+        return dt.hour
+
+
+@dataclass
+class UserHistory:
+    user_id: str
+    events: list[FeedbackEvent] = field(default_factory=list)

ml/agents/tests/test_agents.py

@@ -153,7 +153,7 @@ class TestTimeOfDayAgent:
 
     def test_snapshot_keys(self):
         out = self.agent.compute(_inp())
-        assert {"hour", "day_of_week", "preferred_hour"} == set(out.signals_snapshot)
+        assert {"hour", "day_of_week", "preferred_hour", "quiet_start", "quiet_end"} == set(out.signals_snapshot)
 
 
 # ── RecentPatternsAgent ───────────────────────────────────────────────────────

ml/agents/tests/test_inference.py

@@ -0,0 +1,120 @@
+"""Tests for the inference framework and time-of-day #112 proof."""
+from __future__ import annotations
+
+import sys, os
+sys.path.insert(0, os.path.join(os.path.dirname(__file__), "..", "..", ".."))
+
+import pytest
+from datetime import datetime, timezone
+
+from ml.agents.inference.history import FeedbackEvent, UserHistory
+from ml.agents.inference.framework import run_inference
+from ml.agents.time_of_day import TimeOfDayAgent, MANIFEST as TOD_MANIFEST, MANIFEST
+from ml.agents.base import AgentInput
+
+
+_NOW = datetime(2026, 5, 1, 14, 0, 0, tzinfo=timezone.utc)  # Thursday 14:00
+
+
+def _inp(**kwargs) -> AgentInput:
+    defaults = dict(user_id="u1", tasks=[], profile={}, now=_NOW, agent_prefs={})
+    defaults.update(kwargs)
+    return AgentInput(**defaults)
+
+
+def _event(action: str, hour: int) -> FeedbackEvent:
+    ts = f"2026-05-01T{hour:02d}:00:00+00:00"
+    return FeedbackEvent(action=action, dwell_ms=60_000 if action == "done" else 500, created_at=ts)
+
+
+class TestRunInference:
+    def test_cold_start_when_below_min_history(self):
+        history = UserHistory(user_id="u1", events=[_event("done", 9)] * 5)  # only 5 < 10
+        result = run_inference(TOD_MANIFEST, history)
+        assert result["preferred_hour"] is None  # cold_start_default
+
+    def test_infers_preferred_hour_as_mode(self):
+        # 7 events at 09:00, 3 at 17:00 → preferred_hour should be 9
+        events = [_event("done", 9)] * 7 + [_event("done", 17)] * 3
+        history = UserHistory(user_id="u1", events=events)
+        result = run_inference(TOD_MANIFEST, history)
+        assert result["preferred_hour"] == 9
+
+    def test_infers_preferred_hour_from_majority_hour(self):
+        events = [_event("done", 20)] * 6 + [_event("done", 8)] * 4
+        history = UserHistory(user_id="u1", events=events)
+        result = run_inference(TOD_MANIFEST, history)
+        assert result["preferred_hour"] == 20
+
+    def test_no_inferred_params_returns_empty(self):
+        from ml.agents.manifest import AgentManifest
+        bare = AgentManifest(
+            id="bare", version="1.0.0", description="", pref_schema={},
+            context_schema=[], required_consents=[], output_contract={}, ttl_sec=300,
+        )
+        history = UserHistory(user_id="u1", events=[_event("done", 9)] * 20)
+        result = run_inference(bare, history)
+        assert result == {}
+
+    def test_cold_start_fallback_on_infer_error(self):
+        """infer() raising should fall back to cold_start_default, not crash."""
+        from ml.agents.manifest import InferredParam, AgentManifest
+
+        def _bad_infer(h):
+            raise RuntimeError("oops")
+
+        m = AgentManifest(
+            id="boom", version="1.0.0", description="", pref_schema={},
+            context_schema=[], required_consents=[], output_contract={}, ttl_sec=300,
+            inferred_params=[InferredParam(key="x", ttl_sec=60, cold_start_default=42, min_history=1, infer=_bad_infer)],
+        )
+        history = UserHistory(user_id="u1", events=[_event("done", 9)] * 5)
+        result = run_inference(m, history)
+        assert result["x"] == 42
+
+
+class TestTimeOfDayAgentWithInference:
+    agent = TimeOfDayAgent()
+
+    def test_uses_preferred_hour_from_agent_prefs(self):
+        inp = _inp(agent_prefs={"preferred_hour": 9}, now=datetime(2026, 5, 1, 9, 0, 0, tzinfo=timezone.utc))
+        out = self.agent.compute(inp)
+        assert "peak productivity hour" in out.prompt_text.lower() or "peak" in out.prompt_text
+
+    def test_quiet_window_noon_suppressed(self):
+        inp = _inp(
+            agent_prefs={"quiet_start": "22:00", "quiet_end": "07:00"},
+            now=datetime(2026, 5, 1, 23, 0, 0, tzinfo=timezone.utc),
+        )
+        out = self.agent.compute(inp)
+        assert "quiet window" in out.prompt_text
+
+    def test_quiet_window_not_in_window(self):
+        inp = _inp(
+            agent_prefs={"quiet_start": "22:00", "quiet_end": "07:00"},
+            now=datetime(2026, 5, 1, 14, 0, 0, tzinfo=timezone.utc),
+        )
+        out = self.agent.compute(inp)
+        assert "quiet window" not in out.prompt_text
+
+    def test_agent_prefs_override_profile(self):
+        # agent_prefs.preferred_hour wins over profile.preferred_hour
+        inp = _inp(
+            profile={"preferred_hour": 8},
+            agent_prefs={"preferred_hour": 14},
+            now=datetime(2026, 5, 1, 14, 0, 0, tzinfo=timezone.utc),
+        )
+        out = self.agent.compute(inp)
+        assert "peak productivity hour (14:00)" in out.prompt_text
+
+    def test_no_prefs_falls_back_to_profile(self):
+        inp = _inp(profile={"preferred_hour": 10}, now=datetime(2026, 5, 1, 10, 0, 0, tzinfo=timezone.utc))
+        out = self.agent.compute(inp)
+        assert "peak" in out.prompt_text
+
+    def test_version_bumped(self):
+        assert MANIFEST.version == "1.1.0"
+
+    def test_manifest_has_preferred_hour_param(self):
+        keys = {p.key for p in MANIFEST.inferred_params}
+        assert "preferred_hour" in keys

ml/agents/time_of_day.py

@@ -1,14 +1,26 @@
 from __future__ import annotations
+
+from collections import Counter
 from typing import ClassVar
+
 from .base import BaseAgent, AgentInput, AgentOutput
-from .manifest import AgentManifest
+from .inference.history import UserHistory
+from .manifest import AgentManifest, InferredParam
 
 _DOW_NAMES = ["Monday", "Tuesday", "Wednesday", "Thursday", "Friday", "Saturday", "Sunday"]
 
 
+def _infer_preferred_hour(history: UserHistory) -> int:
+    """Mode hour of day across all 'done' feedback events; falls back to 9."""
+    done_hours = [e.hour for e in history.events if e.action == "done"]
+    if not done_hours:
+        return 9
+    return Counter(done_hours).most_common(1)[0][0]
+
+
 MANIFEST = AgentManifest(
     id="time-of-day",
-    version="1.0.0",
+    version="1.1.0",  # bumped: inferred_params added (ADR-0014 §3, #112)
     description="Frames the current moment relative to the user's productive peak and quiet hours.",
     pref_schema={
         "type": "object",
@@ -30,6 +42,15 @@ MANIFEST = AgentManifest(
     required_consents=["data:core", "agent:time-of-day"],
     output_contract={"type": "snippet", "format": "free_text"},
     ttl_sec=900,
+    inferred_params=[
+        InferredParam(
+            key="preferred_hour",
+            ttl_sec=3_600,       # recompute hourly
+            cold_start_default=None,
+            min_history=10,      # need at least 10 feedback events to be meaningful
+            infer=_infer_preferred_hour,
+        ),
+    ],
 )
 
 
@@ -42,31 +63,63 @@ class TimeOfDayAgent(BaseAgent):
     def compute(self, inp: AgentInput) -> AgentOutput:
         hour = inp.now.hour
         dow = inp.now.weekday()  # 0=Monday … 6=Sunday
-        preferred = inp.profile.get("preferred_hour")
         is_weekend = dow >= 5
 
+        # agent_prefs (inferred or user-set) take precedence over ML profile features.
+        preferred_raw = inp.agent_prefs.get("preferred_hour", inp.profile.get("preferred_hour"))
+        preferred = int(preferred_raw) if preferred_raw is not None else None
+
+        quiet_start: str | None = inp.agent_prefs.get("quiet_start")
+        quiet_end: str | None = inp.agent_prefs.get("quiet_end")
+        in_quiet = self._in_quiet_window(hour, quiet_start, quiet_end)
+
         parts = [f"It is {hour:02d}:00 on {_DOW_NAMES[dow]} ({self._label(hour)})."]
 
         if is_weekend:
             parts.append("Weekend context — prefer personal or reflective tips over work tasks.")
 
+        if in_quiet:
+            parts.append(
+                f"User is in their quiet window ({quiet_start}–{quiet_end}) — "
+                "avoid urgent or demanding tips."
+            )
+
         if preferred is not None:
-            ph = int(preferred)
+            delta = min(abs(hour - preferred), 24 - abs(hour - preferred))
-            delta = min(abs(hour - ph), 24 - abs(hour - ph))  # circular distance
             if delta == 0:
                 parts.append(
-                    f"This is the user's peak productivity hour ({ph:02d}:00) — "
+                    f"This is the user's peak productivity hour ({preferred:02d}:00) — "
-                    f"a high-impact tip is appropriate."
+                    "a high-impact tip is appropriate."
                 )
             elif delta <= 2:
-                parts.append(f"Approaching the user's peak productivity window ({ph:02d}:00).")
+                parts.append(f"Approaching the user's peak productivity window ({preferred:02d}:00).")
         else:
             parts.append("No preferred-hour data yet.")
 
         prompt = " ".join(parts)
-        snapshot = {"hour": hour, "day_of_week": dow, "preferred_hour": preferred}
+        snapshot = {
+            "hour": hour,
+            "day_of_week": dow,
+            "preferred_hour": preferred,
+            "quiet_start": quiet_start,
+            "quiet_end": quiet_end,
+        }
         return self._make_output(inp, prompt, snapshot)
 
+    @staticmethod
+    def _in_quiet_window(hour: int, start: str | None, end: str | None) -> bool:
+        if not start or not end:
+            return False
+        try:
+            sh = int(start.split(":")[0])
+            eh = int(end.split(":")[0])
+        except (ValueError, IndexError):
+            return False
+        if sh <= eh:
+            return sh <= hour < eh
+        # wraps midnight e.g. 22:00–07:00
+        return hour >= sh or hour < eh
+
     @staticmethod
     def _label(hour: int) -> str:
         if 5 <= hour < 12:

ml/serving/main.py

@@ -3,6 +3,7 @@ oO ML Serving — multi-agent orchestrator (ADR-0013).
 
 Contract:
   POST /agents/{agent_id}/compute  run a sub-agent, return prompt snippet
+  POST /agents/{agent_id}/infer    run inference framework for a user, return inferred prefs
   POST /recommend                  orchestrate agent snippets → one tip via LiteLLM
   POST /generate                   LLM tip candidates (legacy; kept for bench/eval)
   GET  /health                     { ok, agents: [...] }
@@ -38,7 +39,8 @@ if _repo_root not in sys.path:
     sys.path.insert(0, _repo_root)
 
 from ml.agents.base import AgentInput  # noqa: E402
-from ml.agents.registry import get_agent, all_agents, all_manifests  # noqa: E402
+from ml.agents.registry import get_agent, all_agents, all_manifests, get_manifest  # noqa: E402
+from ml.agents.inference import run_inference, FeedbackEvent, UserHistory  # noqa: E402
 
 logging_config.configure()
 
@@ -123,6 +125,8 @@ class AgentComputeRequest(BaseModel):
     profile: dict[str, Optional[float]] = {}
     feedback_history: list[dict] = []
     now_iso: Optional[str] = None  # ISO 8601; defaults to utcnow
+    # Per-agent prefs from user_preferences (merged: user source overrides inferred).
+    agent_prefs: dict = {}
 
 
 class AgentComputeResponse(BaseModel):
@@ -135,6 +139,18 @@ class AgentComputeResponse(BaseModel):
     agent_version: str
 
 
+class AgentInferRequest(BaseModel):
+    user_id: str
+    feedback_history: list[dict] = []   # [{action, dwell_ms, created_at}, …]
+
+
+class AgentInferResponse(BaseModel):
+    user_id: str
+    agent_id: str
+    # {key: inferred_value} — caller persists to user_preferences with source='inferred'
+    inferred_prefs: dict
+
+
 class AgentOutputSnippet(BaseModel):
     agent_id: str
     prompt_text: str
@@ -225,6 +241,7 @@ async def compute_agent(agent_id: str, req: AgentComputeRequest) -> AgentCompute
         profile=req.profile,
         feedback_history=req.feedback_history,
         now=now,
+        agent_prefs=req.agent_prefs,
     )
     try:
         output = agent.compute(inp)
@@ -244,6 +261,46 @@ async def compute_agent(agent_id: str, req: AgentComputeRequest) -> AgentCompute
     )
 
 
+@app.post("/agents/{agent_id}/infer", response_model=AgentInferResponse)
+async def infer_agent(agent_id: str, req: AgentInferRequest) -> AgentInferResponse:
+    """Run the inference framework for one agent and return inferred preference values.
+
+    The caller (TS agent-outputs.ts) persists results to user_preferences
+    with source='inferred', skipping keys where source='user' already exists.
+    """
+    try:
+        manifest = get_manifest(agent_id)
+    except KeyError:
+        raise HTTPException(status_code=404, detail=f"Unknown agent: {agent_id!r}")
+
+    if not manifest.inferred_params:
+        return AgentInferResponse(user_id=req.user_id, agent_id=agent_id, inferred_prefs={})
+
+    events = [
+        FeedbackEvent(
+            action=e.get("action", ""),
+            dwell_ms=e.get("dwell_ms"),
+            created_at=e.get("created_at", ""),
+        )
+        for e in req.feedback_history
+    ]
+    history = UserHistory(user_id=req.user_id, events=events)
+
+    t0 = __import__("time").monotonic()
+    inferred = run_inference(manifest, history)
+    latency_ms = round((__import__("time").monotonic() - t0) * 1000, 1)
+
+    log.info(
+        "inference_run",
+        agent_id=agent_id,
+        user_id=req.user_id,
+        n_params=len(inferred),
+        history_len=len(events),
+        latency_ms=latency_ms,
+    )
+    return AgentInferResponse(user_id=req.user_id, agent_id=agent_id, inferred_prefs=inferred)
+
+
 @app.post("/recommend", response_model=RecommendResponse)
 async def recommend(req: RecommendRequest) -> RecommendResponse:
     """Orchestrator: combine pre-computed agent outputs into one tip via LLM.

ml/serving/tests/test_infer_endpoint.py

@@ -0,0 +1,52 @@
+"""POST /agents/{agent_id}/infer — inference framework endpoint."""
+import pytest
+from httpx import AsyncClient, ASGITransport
+
+from main import app
+
+
+@pytest.mark.anyio
+async def test_infer_time_of_day_cold_start():
+    """Fewer than min_history events → cold_start_default for preferred_hour."""
+    transport = ASGITransport(app=app)
+    async with AsyncClient(transport=transport, base_url="http://test") as client:
+        resp = await client.post("/agents/time-of-day/infer", json={
+            "user_id": "u1",
+            "feedback_history": [
+                {"action": "done", "dwell_ms": 60000, "created_at": "2026-05-01T09:00:00+00:00"},
+            ] * 5,  # 5 < min_history=10
+        })
+    assert resp.status_code == 200
+    body = resp.json()
+    assert body["agent_id"] == "time-of-day"
+    assert body["inferred_prefs"]["preferred_hour"] is None
+
+
+@pytest.mark.anyio
+async def test_infer_time_of_day_enough_history():
+    """10+ events → preferred_hour is inferred as the mode done-hour."""
+    events = [{"action": "done", "dwell_ms": 60000, "created_at": "2026-05-01T09:00:00+00:00"}] * 10
+    transport = ASGITransport(app=app)
+    async with AsyncClient(transport=transport, base_url="http://test") as client:
+        resp = await client.post("/agents/time-of-day/infer", json={"user_id": "u1", "feedback_history": events})
+    assert resp.status_code == 200
+    body = resp.json()
+    assert body["inferred_prefs"]["preferred_hour"] == 9
+
+
+@pytest.mark.anyio
+async def test_infer_agent_with_no_inferred_params():
+    """Agents with no inferred_params return an empty dict."""
+    transport = ASGITransport(app=app)
+    async with AsyncClient(transport=transport, base_url="http://test") as client:
+        resp = await client.post("/agents/overdue-task/infer", json={"user_id": "u1", "feedback_history": []})
+    assert resp.status_code == 200
+    assert resp.json()["inferred_prefs"] == {}
+
+
+@pytest.mark.anyio
+async def test_infer_unknown_agent_404():
+    transport = ASGITransport(app=app)
+    async with AsyncClient(transport=transport, base_url="http://test") as client:
+        resp = await client.post("/agents/ghost/infer", json={"user_id": "u1", "feedback_history": []})
+    assert resp.status_code == 404

services/api/src/index.ts

@@ -19,6 +19,7 @@ import { adminRouter, adminInternalRouter } from './routes/admin.js';
 import benchRouter from './routes/bench.js';
 import agentOutputsRouter from './routes/agent-outputs.js';
 import agentRegistryRouter from './routes/agent-registry.js';
+import profileRouter from './routes/profile.js';
 import { mkdir } from 'fs/promises';
 import { dirname } from 'path';
 import { requireAuth } from './middleware/session.js';
@@ -74,6 +75,7 @@ app.use('/api/bench', requireAuth as any, requireAdmin as any, benchRouter);
 // agent-registry mounts first so /registry beats agent-outputs' /:userId pattern.
 app.use('/api/agents', agentRegistryRouter);
 app.use('/api/agents', agentOutputsRouter);
+app.use('/api/profile', profileRouter);
 
 app.use('/api/ml', requireAuth as any, requireAdmin as any, async (req: Request, res: Response) => {
   const mlUrl = config.ML_SERVING_URL;

services/api/src/profile/__tests__/eligibility.test.ts

@@ -0,0 +1,130 @@
+/**
+ * Unit tests for getEligibleAgentIds (ADR-0014 step 5).
+ * DB is mocked via in-memory SQLite; fetchRegistry is mocked per scenario.
+ */
+import { describe, it, expect, vi, beforeAll, beforeEach } from 'vitest';
+import { makeTestDb } from '../../test/db.js';
+import { users, userConsents, userPreferences, userContexts } from '../../db/schema.js';
+
+const testDb = makeTestDb();
+vi.mock('../../db/index.js', () => ({ db: testDb, rawSqlite: testDb.rawSqlite }));
+
+// Registry mock — overridden per test.
+const mockFetchRegistry = vi.fn();
+vi.mock('../../routes/agent-registry.js', () => ({
+  fetchRegistry: (...args: unknown[]) => mockFetchRegistry(...args),
+  _resetRegistryCache: vi.fn(),
+}));
+
+const { getEligibleAgentIds } = await import('../eligibility.js');
+
+const NOW = new Date().toISOString();
+const MANIFEST_DEFAULTS = {
+  version: '1.0.0',
+  description: '',
+  pref_schema: {},
+  context_schema: [],
+  output_contract: {},
+  ttl_sec: 300,
+};
+
+const AGENT_A = { ...MANIFEST_DEFAULTS, id: 'agent-a', required_consents: ['data:core'], silenced_in_contexts: [] };
+const AGENT_B = { ...MANIFEST_DEFAULTS, id: 'agent-b', required_consents: ['data:core', 'data:todoist'], silenced_in_contexts: [] };
+const AGENT_C = { ...MANIFEST_DEFAULTS, id: 'agent-c', required_consents: ['data:core'], silenced_in_contexts: ['vacation'] };
+
+beforeAll(async () => {
+  await testDb.insert(users).values({
+    id: 'u1', email: 'u@test.com', name: null, image: null, role: 'user',
+    consentGiven: false, createdAt: NOW,
+  });
+});
+
+beforeEach(() => {
+  mockFetchRegistry.mockReset();
+});
+
+describe('getEligibleAgentIds', () => {
+  it('returns empty set when registry is unavailable', async () => {
+    mockFetchRegistry.mockRejectedValue(new Error('network'));
+    const ids = await getEligibleAgentIds('u1');
+    expect(ids.size).toBe(0);
+  });
+
+  it('excludes agents whose required consents are not granted', async () => {
+    mockFetchRegistry.mockResolvedValue({ agents: [AGENT_A, AGENT_B] });
+    // only data:core granted
+    await testDb.insert(userConsents).values({ userId: 'u1', consentKey: 'data:core', grantedAt: NOW, revokedAt: null });
+
+    const ids = await getEligibleAgentIds('u1');
+    expect(ids.has('agent-a')).toBe(true);
+    expect(ids.has('agent-b')).toBe(false);
+  });
+
+  it('excludes agents when a required consent is revoked', async () => {
+    mockFetchRegistry.mockResolvedValue({ agents: [AGENT_B] });
+    // grant then revoke data:todoist
+    await testDb.insert(userConsents).values([
+      { userId: 'u1', consentKey: 'data:todoist', grantedAt: NOW, revokedAt: NOW },
+    ]).onConflictDoUpdate({
+      target: [userConsents.userId, userConsents.consentKey],
+      set: { revokedAt: NOW },
+    });
+
+    const ids = await getEligibleAgentIds('u1');
+    expect(ids.has('agent-b')).toBe(false);
+  });
+
+  it('respects legacy consentGiven bit as data:core', async () => {
+    mockFetchRegistry.mockResolvedValue({ agents: [AGENT_A] });
+    // no consent rows, but legacy bit set
+    await testDb.update(users).set({ consentGiven: true });
+
+    const ids = await getEligibleAgentIds('u1');
+    expect(ids.has('agent-a')).toBe(true);
+
+    await testDb.update(users).set({ consentGiven: false });
+  });
+
+  it('silences agents whose silenced_in_contexts intersects active contexts', async () => {
+    mockFetchRegistry.mockResolvedValue({ agents: [AGENT_A, AGENT_C] });
+    // ensure data:core granted
+    await testDb.insert(userConsents).values({ userId: 'u1', consentKey: 'data:core', grantedAt: NOW, revokedAt: null })
+      .onConflictDoUpdate({ target: [userConsents.userId, userConsents.consentKey], set: { revokedAt: null } });
+    // activate vacation context
+    await testDb.insert(userContexts).values({ userId: 'u1', name: 'vacation', active: true, scheduleJson: null, createdAt: NOW });
+
+    const ids = await getEligibleAgentIds('u1');
+    expect(ids.has('agent-a')).toBe(true);
+    expect(ids.has('agent-c')).toBe(false);
+  });
+
+  it('excludes agents explicitly disabled via user_preferences', async () => {
+    mockFetchRegistry.mockResolvedValue({ agents: [AGENT_A] });
+    await testDb.insert(userConsents).values({ userId: 'u1', consentKey: 'data:core', grantedAt: NOW, revokedAt: null })
+      .onConflictDoUpdate({ target: [userConsents.userId, userConsents.consentKey], set: { revokedAt: null } });
+    await testDb.insert(userPreferences).values({
+      userId: 'u1', scope: 'agent:agent-a', key: 'enabled', valueJson: 'false', source: 'user', updatedAt: NOW,
+    }).onConflictDoUpdate({
+      target: [userPreferences.userId, userPreferences.scope, userPreferences.key],
+      set: { valueJson: 'false' },
+    });
+
+    const ids = await getEligibleAgentIds('u1');
+    expect(ids.has('agent-a')).toBe(false);
+  });
+
+  it('includes agents when enabled pref is true (or absent)', async () => {
+    mockFetchRegistry.mockResolvedValue({ agents: [AGENT_A] });
+    await testDb.insert(userConsents).values({ userId: 'u1', consentKey: 'data:core', grantedAt: NOW, revokedAt: null })
+      .onConflictDoUpdate({ target: [userConsents.userId, userConsents.consentKey], set: { revokedAt: null } });
+    await testDb.insert(userPreferences).values({
+      userId: 'u1', scope: 'agent:agent-a', key: 'enabled', valueJson: 'true', source: 'user', updatedAt: NOW,
+    }).onConflictDoUpdate({
+      target: [userPreferences.userId, userPreferences.scope, userPreferences.key],
+      set: { valueJson: 'true' },
+    });
+
+    const ids = await getEligibleAgentIds('u1');
+    expect(ids.has('agent-a')).toBe(true);
+  });
+});

services/api/src/profile/eligibility.ts

@@ -0,0 +1,88 @@
+/**
+ * Registry-driven agent eligibility filter (ADR-0014 step 5).
+ *
+ * Rules (all must pass for an agent to be eligible):
+ *   1. All required_consents are granted and not revoked.
+ *   2. No silenced_in_contexts entry matches an active context.
+ *   3. user_preferences[scope='agent:<id>', key='enabled'] is not false.
+ *
+ * Fail-closed: if the registry is unavailable, returns an empty set so the
+ * orchestrator falls back to the random policy rather than proceeding without
+ * consent checks.
+ */
+import { db } from '../db/index.js';
+import { users, userConsents, userPreferences, userContexts } from '../db/schema.js';
+import { eq, and, isNull } from 'drizzle-orm';
+import { fetchRegistry } from '../routes/agent-registry.js';
+
+export interface AgentManifestWire {
+  id: string;
+  required_consents: string[];
+  silenced_in_contexts: string[];
+  [key: string]: unknown;
+}
+
+interface RegistryPayload {
+  agents: AgentManifestWire[];
+}
+
+export async function getEligibleAgentIds(userId: string): Promise<Set<string>> {
+  let registry: RegistryPayload;
+  try {
+    registry = (await fetchRegistry()) as RegistryPayload;
+  } catch {
+    return new Set();
+  }
+
+  const [consentRows, prefRows, contextRows, userRow] = await Promise.all([
+    db
+      .select({ consentKey: userConsents.consentKey })
+      .from(userConsents)
+      .where(and(eq(userConsents.userId, userId), isNull(userConsents.revokedAt))),
+    db
+      .select({ scope: userPreferences.scope, key: userPreferences.key, valueJson: userPreferences.valueJson })
+      .from(userPreferences)
+      .where(eq(userPreferences.userId, userId)),
+    db
+      .select({ name: userContexts.name, active: userContexts.active })
+      .from(userContexts)
+      .where(and(eq(userContexts.userId, userId), eq(userContexts.active, true))),
+    db
+      .select({ consentGiven: users.consentGiven })
+      .from(users)
+      .where(eq(users.id, userId))
+      .limit(1),
+  ]);
+
+  // Active consents (granted + not revoked)
+  const activeConsents = new Set(consentRows.map((r) => r.consentKey));
+  // Legacy fallback: consentGiven bit counts as data:core
+  if (!activeConsents.has('data:core') && userRow[0]?.consentGiven) {
+    activeConsents.add('data:core');
+  }
+
+  // Active context names
+  const activeContextNames = new Set(contextRows.map((r) => r.name));
+
+  // Per-agent enabled flag from user_preferences
+  const agentEnabled: Record<string, boolean> = {};
+  for (const p of prefRows) {
+    if (!p.scope.startsWith('agent:')) continue;
+    if (p.key !== 'enabled') continue;
+    try {
+      agentEnabled[p.scope] = JSON.parse(p.valueJson) as boolean;
+    } catch {
+      // ignore malformed
+    }
+  }
+
+  const eligible = new Set<string>();
+  for (const manifest of registry.agents) {
+    if (!manifest.required_consents.every((c) => activeConsents.has(c))) continue;
+    if (manifest.silenced_in_contexts.some((ctx) => activeContextNames.has(ctx))) continue;
+    const enabledPref = agentEnabled[`agent:${manifest.id}`];
+    if (enabledPref === false) continue;
+    eligible.add(manifest.id);
+  }
+  return eligible;
+}

services/api/src/routes/__tests__/profile.test.ts

@@ -0,0 +1,201 @@
+/**
+ * Integration tests for GET/PATCH /api/profile (ADR-0014 step 4).
+ * Real in-memory SQLite; auth middleware mocked so requests arrive as 'user-1'.
+ */
+import { describe, it, expect, vi, beforeAll, afterAll } from 'vitest';
+import express from 'express';
+import * as http from 'http';
+import { makeTestDb } from '../../test/db.js';
+import { users, userPreferences, userConsents, userContexts } from '../../db/schema.js';
+
+const testDb = makeTestDb();
+
+vi.mock('../../db/index.js', () => ({ db: testDb, rawSqlite: testDb.rawSqlite }));
+
+vi.mock('../../middleware/session.js', () => ({
+  sessionMiddleware: (_req: express.Request, _res: express.Response, next: express.NextFunction) =>
+    next(),
+  requireAuth: (req: express.Request, _res: express.Response, next: express.NextFunction) => {
+    (req as any).userId = 'user-1';
+    next();
+  },
+}));
+
+function call(
+  server: http.Server,
+  method: string,
+  path: string,
+  body?: unknown,
+): Promise<{ status: number; body: unknown }> {
+  return new Promise((resolve, reject) => {
+    const { port } = server.address() as { port: number };
+    const req = http.request(
+      { method, hostname: '127.0.0.1', port, path, headers: { 'Content-Type': 'application/json' } },
+      (res) => {
+        let data = '';
+        res.on('data', (c) => (data += c));
+        res.on('end', () => {
+          try { resolve({ status: res.statusCode!, body: JSON.parse(data) }); }
+          catch { resolve({ status: res.statusCode!, body: data }); }
+        });
+      },
+    );
+    req.on('error', reject);
+    if (body !== undefined) req.write(JSON.stringify(body));
+    req.end();
+  });
+}
+
+function startServer(app: express.Application): Promise<{ server: http.Server; call: (method: string, path: string, body?: unknown) => ReturnType<typeof call> }> {
+  return new Promise((resolve) => {
+    const server = http.createServer(app);
+    server.listen(0, () =>
+      resolve({ server, call: (m, p, b) => call(server, m, p, b) }),
+    );
+  });
+}
+
+const profileRouter = (await import('../profile.js')).default;
+const app = express();
+app.use(express.json());
+app.use('/api/profile', profileRouter);
+
+const { server, call: c } = await startServer(app);
+afterAll(() => server.close());
+
+const NOW = new Date().toISOString();
+
+beforeAll(async () => {
+  await testDb.insert(users).values({
+    id: 'user-1',
+    email: 'a@example.com',
+    name: 'Alice',
+    image: null,
+    role: 'user',
+    consentGiven: false,
+    tone: 'direct',
+    tipKindsJson: JSON.stringify(['task', 'advice']),
+    createdAt: NOW,
+  });
+});
+
+describe('GET /api/profile', () => {
+  it('returns user globals with empty prefs/consents/contexts', async () => {
+    const res = await c('GET', '/api/profile');
+    expect(res.status).toBe(200);
+    const body = res.body as any;
+    expect(body.user).toMatchObject({ id: 'user-1', tone: 'direct', tipKinds: ['task', 'advice'] });
+    expect(body.prefs).toEqual({});
+    expect(body.consents).toEqual({});
+    expect(body.contexts).toEqual([]);
+  });
+
+  it('surfaces legacy consentGiven as data:core when no consent row exists', async () => {
+    await testDb.update(users).set({ consentGiven: true, consentAt: NOW });
+    const res = await c('GET', '/api/profile');
+    expect((res.body as any).consents['data:core']).toMatchObject({ revokedAt: null });
+    await testDb.update(users).set({ consentGiven: false });
+  });
+
+  it('includes prefs grouped by scope', async () => {
+    await testDb.insert(userPreferences).values([
+      { userId: 'user-1', scope: 'orchestrator', key: 'quietHours', valueJson: '"22:00-07:00"', source: 'user', updatedAt: NOW },
+      { userId: 'user-1', scope: 'agent:focus-area', key: 'areas', valueJson: '["work","health"]', source: 'inferred', updatedAt: NOW },
+    ]);
+    const res = await c('GET', '/api/profile');
+    const body = res.body as any;
+    expect(body.prefs['orchestrator']).toMatchObject({ quietHours: '22:00-07:00' });
+    expect(body.prefs['agent:focus-area']).toMatchObject({ areas: ['work', 'health'] });
+  });
+
+  it('includes consents', async () => {
+    await testDb.insert(userConsents).values([
+      { userId: 'user-1', consentKey: 'data:core', grantedAt: NOW, revokedAt: null },
+      { userId: 'user-1', consentKey: 'data:todoist', grantedAt: NOW, revokedAt: NOW },
+    ]);
+    const body = (await c('GET', '/api/profile')).body as any;
+    expect(body.consents['data:core'].revokedAt).toBeNull();
+    expect(body.consents['data:todoist'].revokedAt).toBe(NOW);
+  });
+
+  it('includes contexts', async () => {
+    await testDb.insert(userContexts).values({
+      userId: 'user-1', name: 'work', active: true, scheduleJson: null, createdAt: NOW,
+    });
+    const body = (await c('GET', '/api/profile')).body as any;
+    expect(body.contexts).toContainEqual(expect.objectContaining({ name: 'work', active: true }));
+  });
+});
+
+describe('PATCH /api/profile/prefs/:scope', () => {
+  it('upserts preference keys with source=user', async () => {
+    const res = await c('PATCH', '/api/profile/prefs/orchestrator', { tone: 'gentle' });
+    expect(res.status).toBe(200);
+    expect(res.body).toEqual({ ok: true });
+
+    const body = (await c('GET', '/api/profile')).body as any;
+    expect(body.prefs['orchestrator']['tone']).toBe('gentle');
+  });
+
+  it('overwrites an inferred value with user source', async () => {
+    await testDb.insert(userPreferences).values({
+      userId: 'user-1', scope: 'agent:momentum', key: 'enabled', valueJson: 'false',
+      source: 'inferred', updatedAt: NOW,
+    }).onConflictDoUpdate({
+      target: [userPreferences.userId, userPreferences.scope, userPreferences.key],
+      set: { valueJson: 'false', source: 'inferred', updatedAt: NOW },
+    });
+
+    await c('PATCH', '/api/profile/prefs/agent:momentum', { enabled: true });
+    const body = (await c('GET', '/api/profile')).body as any;
+    expect(body.prefs['agent:momentum']['enabled']).toBe(true);
+  });
+
+  it('returns 400 for non-object body', async () => {
+    const res = await c('PATCH', '/api/profile/prefs/orchestrator', [1, 2]);
+    expect(res.status).toBe(400);
+  });
+});
+
+describe('PATCH /api/profile/consents', () => {
+  it('grants a new consent key', async () => {
+    const res = await c('PATCH', '/api/profile/consents', { grant: ['data:calendar'] });
+    expect(res.status).toBe(200);
+    const body = (await c('GET', '/api/profile')).body as any;
+    expect(body.consents['data:calendar'].revokedAt).toBeNull();
+  });
+
+  it('revokes an existing active consent', async () => {
+    await c('PATCH', '/api/profile/consents', { grant: ['agent:overdue-task'] });
+    await c('PATCH', '/api/profile/consents', { revoke: ['agent:overdue-task'] });
+    const body = (await c('GET', '/api/profile')).body as any;
+    expect(body.consents['agent:overdue-task'].revokedAt).not.toBeNull();
+  });
+
+  it('returns 400 when grant is not an array', async () => {
+    const res = await c('PATCH', '/api/profile/consents', { grant: 'data:core' });
+    expect(res.status).toBe(400);
+  });
+});
+
+describe('PATCH /api/profile/contexts', () => {
+  it('creates a new context', async () => {
+    const res = await c('PATCH', '/api/profile/contexts', { name: 'vacation', active: false });
+    expect(res.status).toBe(200);
+    const body = (await c('GET', '/api/profile')).body as any;
+    expect(body.contexts).toContainEqual(expect.objectContaining({ name: 'vacation', active: false }));
+  });
+
+  it('toggles active on existing context', async () => {
+    await c('PATCH', '/api/profile/contexts', { name: 'home', active: false });
+    await c('PATCH', '/api/profile/contexts', { name: 'home', active: true });
+    const body = (await c('GET', '/api/profile')).body as any;
+    const ctx = (body.contexts as any[]).find((x) => x.name === 'home');
+    expect(ctx?.active).toBe(true);
+  });
+
+  it('returns 400 when name is missing', async () => {
+    const res = await c('PATCH', '/api/profile/contexts', { active: true });
+    expect(res.status).toBe(400);
+  });
+});

services/api/src/routes/__tests__/recommender.test.ts

@@ -13,7 +13,8 @@ import { describe, it, expect, vi, beforeAll, afterEach } from 'vitest';
 import express from 'express';
 import * as http from 'http';
 import { makeTestDb } from '../../test/db.js';
-import { users, integrationTokens, tipScores } from '../../db/schema.js';
+import { users, integrationTokens, tipScores, agentOutputs, userConsents } from '../../db/schema.js';
+import { nanoid } from 'nanoid';
 
 const testDb = makeTestDb();
 
@@ -155,4 +156,77 @@ describe('POST /recommend integration', () => {
     expect(row.promptVersion).toBeNull();
     expect(row.llmModel).toBeNull();
   });
+
+  it('eligibility filter: only passes consented agent outputs to ml/serving', async () => {
+    const NOW = new Date().toISOString();
+    const FUTURE = new Date(Date.now() + 60_000).toISOString();
+
+    // Grant data:core only — not data:todoist
+    await testDb.insert(userConsents).values([
+      { userId: 'user-1', consentKey: 'data:core', grantedAt: NOW, revokedAt: null },
+    ]).onConflictDoUpdate({
+      target: [userConsents.userId, userConsents.consentKey],
+      set: { revokedAt: null },
+    });
+
+    // Two agent outputs: time-of-day (needs data:core only) and overdue-task (needs data:todoist too)
+    await testDb.insert(agentOutputs).values([
+      {
+        id: nanoid(), userId: 'user-1', agentId: 'time-of-day',
+        promptText: 'It is morning.',
+        computedAt: NOW, expiresAt: FUTURE, agentVersion: '1.0.0',
+      },
+      {
+        id: nanoid(), userId: 'user-1', agentId: 'overdue-task',
+        promptText: 'You have overdue tasks.',
+        computedAt: NOW, expiresAt: FUTURE, agentVersion: '1.0.0',
+      },
+    ]);
+
+    // Manifest: time-of-day requires ['data:core'], overdue-task requires ['data:core','data:todoist']
+    const registry = {
+      agents: [
+        { id: 'time-of-day', required_consents: ['data:core'], silenced_in_contexts: [], version: '1.0.0', description: '', pref_schema: {}, context_schema: [], output_contract: {}, ttl_sec: 300, inferred_params: [] },
+        { id: 'overdue-task', required_consents: ['data:core', 'data:todoist'], silenced_in_contexts: [], version: '1.0.0', description: '', pref_schema: {}, context_schema: [], output_contract: {}, ttl_sec: 300, inferred_params: [] },
+      ],
+    };
+
+    let capturedAgentOutputs: { agent_id: string }[] = [];
+    globalThis.fetch = vi.fn().mockImplementation((url: string) => {
+      const u = String(url);
+      if (u.includes('todoist.com')) {
+        return Promise.resolve({ ok: true, status: 200, json: async () => ({ results: [] }) } as any);
+      }
+      if (u.includes('/agents/registry')) {
+        return Promise.resolve({ ok: true, status: 200, json: async () => registry } as any);
+      }
+      if (u.includes('/recommend')) {
+        return Promise.resolve({
+          ok: true, status: 200,
+          json: async (req?: Request) => {
+            // The body has already been sent; capture via the mock call args instead
+            return { tip: { id: 'tip-x', content: 'Stay focused.' }, model: 'tip-generator' };
+          },
+        } as any);
+      }
+      return Promise.resolve({ ok: false, status: 500 } as any);
+    });
+
+    // Intercept the /recommend body to inspect what agent_outputs were sent
+    const origFetch = globalThis.fetch as ReturnType<typeof vi.fn>;
+    const wrappedFetch = vi.fn().mockImplementation(async (url: string, init?: RequestInit) => {
+      if (String(url).includes('/recommend') && init?.body) {
+        const body = JSON.parse(init.body as string);
+        capturedAgentOutputs = body.agent_outputs ?? [];
+      }
+      return origFetch(url, init);
+    });
+    globalThis.fetch = wrappedFetch;
+
+    const { status } = await post(`${baseUrl}/api/recommend`);
+    expect(status).toBe(200);
+
+    // Only time-of-day should have been passed; overdue-task is blocked (missing data:todoist)
+    expect(capturedAgentOutputs.map((a) => a.agent_id)).toEqual(['time-of-day']);
+  });
 });

services/api/src/routes/agent-outputs.ts

@@ -1,7 +1,7 @@
 import { Router, type Request, type Response, type IRouter } from 'express';
 import { nanoid } from 'nanoid';
 import { db } from '../db/index.js';
-import { agentOutputs, tipFeedback, tipViews } from '../db/schema.js';
+import { agentOutputs, tipFeedback, tipViews, userPreferences } from '../db/schema.js';
 import { eq, and, gt, lt } from 'drizzle-orm';
 import { config } from '../config.js';
 import { getProfile, type Profile } from '../profile/builder.js';
@@ -78,6 +78,54 @@ router.get('/active-users', async (req: Request, res: Response) => {
 
 // ── Core compute logic (used by route + scheduler) ───────────────────────────
 
+/** Load agent prefs for a user from user_preferences, merging user+inferred.
+ * User source wins: if both exist, the 'user' row is returned. */
+async function loadAgentPrefs(userId: string, agentId: string): Promise<Record<string, unknown>> {
+  const scope = `agent:${agentId}`;
+  const rows = await db
+    .select({ key: userPreferences.key, valueJson: userPreferences.valueJson, source: userPreferences.source })
+    .from(userPreferences)
+    .where(and(eq(userPreferences.userId, userId), eq(userPreferences.scope, scope)));
+
+  // Build merged dict: 'user' source takes precedence over 'inferred'
+  const merged: Record<string, { value: unknown; source: string }> = {};
+  for (const row of rows) {
+    try {
+      const value = JSON.parse(row.valueJson);
+      const existing = merged[row.key];
+      if (!existing || row.source === 'user') {
+        merged[row.key] = { value, source: row.source };
+      }
+    } catch {
+      // skip malformed
+    }
+  }
+  return Object.fromEntries(Object.entries(merged).map(([k, v]) => [k, v.value]));
+}
+
+/** Persist inferred prefs to user_preferences, skipping keys the user has explicitly set. */
+async function persistInferredPrefs(
+  userId: string,
+  agentId: string,
+  inferredPrefs: Record<string, unknown>,
+): Promise<void> {
+  if (!Object.keys(inferredPrefs).length) return;
+  const scope = `agent:${agentId}`;
+  const now = new Date().toISOString();
+  for (const [key, value] of Object.entries(inferredPrefs)) {
+    const valueJson = JSON.stringify(value);
+    await db
+      .insert(userPreferences)
+      .values({ userId, scope, key, valueJson, source: 'inferred', updatedAt: now })
+      .onConflictDoUpdate({
+        target: [userPreferences.userId, userPreferences.scope, userPreferences.key],
+        set: { valueJson, updatedAt: now },
+        // Only overwrite rows already marked inferred; user overrides are untouched.
+        setWhere: eq(userPreferences.source, 'inferred'),
+      });
+  }
+}
+
 export async function computeAndStore(userId: string, agentId: string): Promise<void> {
   let tasks: object[] = [];
   try {
@@ -111,10 +159,13 @@ export async function computeAndStore(userId: string, agentId: string): Promise<
     created_at: f.createdAt,
   }));
 
+  // Load agent prefs (user overrides + previous inferences) to inject into the compute call.
+  const agentPrefs = await loadAgentPrefs(userId, agentId);
+
   const mlResp = await fetch(`${config.ML_SERVING_URL}/agents/${agentId}/compute`, {
     method: 'POST',
     headers: { 'Content-Type': 'application/json' },
-    body: JSON.stringify({ user_id: userId, tasks, profile, feedback_history: feedbackHistory }),
+    body: JSON.stringify({ user_id: userId, tasks, profile, feedback_history: feedbackHistory, agent_prefs: agentPrefs }),
     signal: AbortSignal.timeout(15_000),
   });
 
@@ -129,6 +180,23 @@ export async function computeAndStore(userId: string, agentId: string): Promise<
   };
 
   await storeAgentOutput(output);
+
+  // Run inference framework for this agent and persist results.
+  // Failures are non-fatal — the compute result is already stored.
+  try {
+    const inferResp = await fetch(`${config.ML_SERVING_URL}/agents/${agentId}/infer`, {
+      method: 'POST',
+      headers: { 'Content-Type': 'application/json' },
+      body: JSON.stringify({ user_id: userId, feedback_history: feedbackHistory }),
+      signal: AbortSignal.timeout(10_000),
+    });
+    if (inferResp.ok) {
+      const inferResult = await inferResp.json() as { inferred_prefs: Record<string, unknown> };
+      await persistInferredPrefs(userId, agentId, inferResult.inferred_prefs);
+    }
+  } catch {
+    // inference failure is non-fatal
+  }
 }
 
 // ── POST /api/agents/:agentId/compute ─────────────────────────────────────────

services/api/src/routes/agent-registry.ts

@@ -13,7 +13,7 @@ export function _resetRegistryCache() {
   _cache = null;
 }
 
-async function fetchRegistry(): Promise<unknown> {
+export async function fetchRegistry(): Promise<unknown> {
   if (_cache && Date.now() - _cache.fetchedAt < CACHE_TTL_MS) return _cache.payload;
   const upstream = await fetch(`${config.ML_SERVING_URL}/agents/registry`, {
     signal: AbortSignal.timeout(5000),

services/api/src/routes/profile.ts

@@ -0,0 +1,202 @@
+/**
+ * GET  /api/profile          — read-through: user globals + prefs + contexts + consents
+ * PATCH /api/profile/prefs/:scope — upsert user_preferences rows (source='user')
+ * PATCH /api/profile/consents    — grant or revoke consent keys
+ * PATCH /api/profile/contexts    — activate/deactivate or create user contexts
+ *
+ * ADR-0014 step 4.
+ */
+import { Router, type Response, type IRouter } from 'express';
+import { db } from '../db/index.js';
+import {
+  users,
+  userPreferences,
+  userConsents,
+  userContexts,
+} from '../db/schema.js';
+import { eq, and, isNull } from 'drizzle-orm';
+import { requireAuth, type AuthenticatedRequest } from '../middleware/session.js';
+
+const router: IRouter = Router();
+
+// ── GET /api/profile ─────────────────────────────────────────────────────────
+
+router.get('/', requireAuth as any, async (req: AuthenticatedRequest, res: Response) => {
+  const userId = req.userId!;
+
+  const [user] = await db.select().from(users).where(eq(users.id, userId)).limit(1);
+  if (!user || user.deletedAt) {
+    res.status(404).json({ error: 'User not found' });
+    return;
+  }
+
+  const [prefs, consents, contexts] = await Promise.all([
+    db.select().from(userPreferences).where(eq(userPreferences.userId, userId)),
+    db.select().from(userConsents).where(eq(userConsents.userId, userId)),
+    db.select().from(userContexts).where(eq(userContexts.userId, userId)),
+  ]);
+
+  // Group prefs by scope: { 'orchestrator': { key: value_json, … }, 'agent:foo': { … } }
+  const prefsByScope: Record<string, Record<string, unknown>> = {};
+  for (const p of prefs) {
+    if (!prefsByScope[p.scope]) prefsByScope[p.scope] = {};
+    try {
+      prefsByScope[p.scope][p.key] = JSON.parse(p.valueJson);
+    } catch {
+      prefsByScope[p.scope][p.key] = p.valueJson;
+    }
+  }
+
+  // Consents: include both active and revoked (callers can filter on revokedAt)
+  // Also fold in the legacy consentGiven bit if no user_consents row exists yet.
+  const consentMap: Record<string, { grantedAt: string; revokedAt: string | null }> = {};
+  for (const c of consents) {
+    consentMap[c.consentKey] = { grantedAt: c.grantedAt, revokedAt: c.revokedAt ?? null };
+  }
+  // Legacy fallback: if data:core is missing and the old bit is set, synthesise it.
+  if (!consentMap['data:core'] && user.consentGiven) {
+    consentMap['data:core'] = { grantedAt: user.consentAt ?? user.createdAt, revokedAt: null };
+  }
+
+  res.json({
+    user: {
+      id: user.id,
+      email: user.email,
+      name: user.name,
+      image: user.image,
+      tone: user.tone ?? null,
+      tipKinds: user.tipKindsJson ? JSON.parse(user.tipKindsJson) : null,
+    },
+    prefs: prefsByScope,
+    consents: consentMap,
+    contexts: contexts.map((c) => ({
+      name: c.name,
+      active: c.active,
+      schedule: c.scheduleJson ? JSON.parse(c.scheduleJson) : null,
+      createdAt: c.createdAt,
+    })),
+  });
+});
+
+// ── PATCH /api/profile/prefs/:scope ──────────────────────────────────────────
+// Body: { [key]: value }  — each key is upserted as source='user'.
+
+router.patch('/prefs/:scope', requireAuth as any, async (req: AuthenticatedRequest, res: Response) => {
+  const userId = req.userId!;
+  const { scope } = req.params;
+  const body = req.body as Record<string, unknown>;
+
+  if (!scope || typeof scope !== 'string') {
+    res.status(400).json({ error: 'scope is required' });
+    return;
+  }
+  if (!body || typeof body !== 'object' || Array.isArray(body)) {
+    res.status(400).json({ error: 'body must be a JSON object' });
+    return;
+  }
+
+  const now = new Date().toISOString();
+  for (const [key, value] of Object.entries(body)) {
+    const valueJson = JSON.stringify(value);
+    await db
+      .insert(userPreferences)
+      .values({ userId, scope, key, valueJson, source: 'user', updatedAt: now })
+      .onConflictDoUpdate({
+        target: [userPreferences.userId, userPreferences.scope, userPreferences.key],
+        set: { valueJson, source: 'user', updatedAt: now },
+      });
+  }
+
+  res.json({ ok: true });
+});
+
+// ── PATCH /api/profile/consents ───────────────────────────────────────────────
+// Body: { grant?: string[], revoke?: string[] }
+
+router.patch('/consents', requireAuth as any, async (req: AuthenticatedRequest, res: Response) => {
+  const userId = req.userId!;
+  const { grant = [], revoke = [] } = req.body as { grant?: string[]; revoke?: string[] };
+
+  if (!Array.isArray(grant) || !Array.isArray(revoke)) {
+    res.status(400).json({ error: 'grant and revoke must be arrays' });
+    return;
+  }
+
+  const now = new Date().toISOString();
+
+  for (const key of grant) {
+    await db
+      .insert(userConsents)
+      .values({ userId, consentKey: key, grantedAt: now, revokedAt: null })
+      .onConflictDoUpdate({
+        target: [userConsents.userId, userConsents.consentKey],
+        set: { grantedAt: now, revokedAt: null },
+      });
+  }
+
+  for (const key of revoke) {
+    await db
+      .update(userConsents)
+      .set({ revokedAt: now })
+      .where(
+        and(
+          eq(userConsents.userId, userId),
+          eq(userConsents.consentKey, key),
+          isNull(userConsents.revokedAt),
+        ),
+      );
+  }
+
+  res.json({ ok: true });
+});
+
+// ── PATCH /api/profile/contexts ───────────────────────────────────────────────
+// Body: { name: string, active?: boolean, schedule?: object|null }
+// Creates the row if it doesn't exist; toggles active / updates schedule.
+
+router.patch('/contexts', requireAuth as any, async (req: AuthenticatedRequest, res: Response) => {
+  const userId = req.userId!;
+  const { name, active, schedule } = req.body as {
+    name?: string;
+    active?: boolean;
+    schedule?: unknown;
+  };
+
+  if (!name || typeof name !== 'string') {
+    res.status(400).json({ error: 'name is required' });
+    return;
+  }
+
+  const now = new Date().toISOString();
+  const scheduleJson = schedule !== undefined ? JSON.stringify(schedule) : undefined;
+
+  const existing = await db
+    .select()
+    .from(userContexts)
+    .where(and(eq(userContexts.userId, userId), eq(userContexts.name, name)))
+    .limit(1);
+
+  if (existing.length === 0) {
+    await db.insert(userContexts).values({
+      userId,
+      name,
+      active: active ?? false,
+      scheduleJson: scheduleJson ?? null,
+      createdAt: now,
+    });
+  } else {
+    const set: Partial<typeof userContexts.$inferInsert> = {};
+    if (active !== undefined) set.active = active;
+    if (scheduleJson !== undefined) set.scheduleJson = scheduleJson;
+    if (Object.keys(set).length > 0) {
+      await db
+        .update(userContexts)
+        .set(set)
+        .where(and(eq(userContexts.userId, userId), eq(userContexts.name, name)));
+    }
+  }
+
+  res.json({ ok: true });
+});
+
+export default router;

services/api/src/routes/recommender.ts

@@ -12,6 +12,7 @@ import { todoistSource, dueAgeDays } from '../signals/todoist.js';
 export { dueAgeDays };
 import { SignalAggregator } from '../signals/aggregator.js';
 import { getActiveAgentOutputs } from './agent-outputs.js';
+import { getEligibleAgentIds } from '../profile/eligibility.js';
 
 const router: ExpressRouter = Router();
 
@@ -58,11 +59,13 @@ async function fetchOrchestratorTip(
   dayOfWeek: number,
   traceparent?: string,
 ): Promise<OrchestratorResult | null> {
-  const agentRows = await getActiveAgentOutputs(userId);
+  const [allAgentRows, eligibleIds] = await Promise.all([
-  const agentOutputs = agentRows.map((r) => ({
+    getActiveAgentOutputs(userId),
-    agent_id: r.agentId,
+    getEligibleAgentIds(userId),
-    prompt_text: r.promptText,
+  ]);
-  }));
+  const agentOutputs = allAgentRows
+    .filter((r) => eligibleIds.has(r.agentId))
+    .map((r) => ({ agent_id: r.agentId, prompt_text: r.promptText }));
 
   const tasks = signals.slice(0, 10).map((s) => ({
     content: s.content,