docs(readme): fix four pieces of drift from the rest of the README

- Quick start: replaced "Phase 4 work" (internal jargon) with
  "auto-supervision not yet implemented" — readers landing here
  shouldn't need to know our phase numbering.
- "Where your memories live" table: added the ``cost.json`` row
  (was already in the deeper Storage-on-disk tree but missing from
  the user-facing quickstart summary).
- Layout section: "177 pytest tests" → pointer to the Status
  section's current count, so this stops drifting every time we
  add tests.
- Cost section: ambient priming overhead was listed as "~500
  chars / ~150 tokens" with "BM25 + embeddings" pipeline. Updated
  to "~1500 chars / ~400 tokens" (matches the actual Tier-1
  budget) and added the cross-encoder rerank to the pipeline
  description (matches the Tier-1 table and what PR #5 actually
  shipped).

Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>

Author: nickroci

README.md

@@ -107,7 +107,8 @@ cd ../tools/search && uv sync
 #    machine — it listens on ~/.agent-mem/priming.sock and answers Tier-1
 #    priming requests from every hook on every project.
 cd /path/to/ultan/daemon && uv run agent-mem-daemon -v
-#    (nohup, tmux, or a launchd plist if you want it persistent — Phase 4 work.)
+#    (nohup, tmux, or a launchd plist if you want it persistent — auto-supervision
+#     not yet implemented.)
 
 # 5. Open Claude Code in any project (no per-project setup needed once the
 #    hooks are global) and work normally. Entries land under
@@ -126,6 +127,7 @@ Everything Ultan owns lives under **`~/.agent-mem/`** on your local disk — no
 | `~/.agent-mem/.bm25.idx`, `.embeddings.idx` | Search indexes over the library. Rebuilt automatically when the library changes. |
 | `~/.agent-mem/sweep-state.json` | Last-decay-sweep timestamp (24h cooldown). |
 | `~/.agent-mem/pending-nudges.md` | Scholar writes nudges here; the hook reads and clears them on the next turn. |
+| `~/.agent-mem/cost.json` | Running tally of LLM spend across Librarian / Scholar / Advisor calls. |
 | `~/.agent-mem/runs/` | Per-call audit log (cost, duration, decisions) + full LLM transcripts (7-day TTL). |
 
 See *Storage on disk* below for the full layout including the folders inside `knowledge/`.
@@ -234,7 +236,7 @@ agent-mem/
   README.md             ← this file
   daemon/               ← the long-lived event-ingest daemon
     agent_mem_daemon/   ← package
-    tests/              ← 177 pytest tests
+    tests/              ← pytest suite (see Status below for current count)
     pyproject.toml      ← uv-managed
   src/                  ← Phase-0 hook layer (forked from claude-memory-compiler)
     hooks/              ← UserPromptSubmit, PostToolUse, Stop, ...
@@ -278,7 +280,7 @@ Storage on disk:
 
 - **Librarian (Sonnet)** runs after each session's quiet period (per-session debounce, default 30s). With moderate activity that's ~10-30 invocations per working day per project. Each is a few thousand input tokens (prompt + library snapshot + buffer) plus a few hundred output tokens.
 - **Scholar (Opus)** runs in batches — every 3 Librarian packets or 60s, whichever first. Each batch is one Opus call: prompt + accumulated proposals, ~30s wall time, ~$0.20-0.50 per batch on pay-as-you-go pricing.
-- **Ambient priming (Tier 1)** is daemon-side BM25 + embeddings, **no LLM cost**, but it injects up to 500 chars into every UserPromptSubmit — call it ~150 tokens of prompt overhead per turn.
+- **Ambient priming (Tier 1)** is daemon-side BM25 + embeddings + cross-encoder rerank, **no LLM cost**, but it injects up to 1500 chars into every UserPromptSubmit — call it ~400 tokens of prompt overhead per turn.
 - **Advisor (`/ultan-advisor`)** is one Sonnet call (Librarian step) + one Opus call (Scholar synthesis) per invocation. ~$0.30-0.50 each.
 - **PreToolUse Tier 3** is pure deterministic regex match, **no LLM cost**, sub-100ms.