chore(config): allow read-only file access for teammates + system-economic-specialization workshop brief

- settings.json: add Read(repo/**), Read(/tmp/**), Grep, Glob to the
  allowlist. Teammates inherit settings.json (not settings.local.json),
  which had no Read rule — so every teammate file read escalated to the
  lead as a per-file prompt. This closes that gap.
- Add the system-economic-specialization workshop brief (lore-authored
  per-system dominant-commodity field + within-system weighted variation;
  Q5 covers adding a small identity field set in the same 300-system pass).

Note: agent memory-strip (6c93b698f) + this read-allow reach teammates
only after a Claude Code restart (settings/agent-def snapshot is taken at
lead startup).

Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>

.claude/settings.json

@@ -63,6 +63,12 @@
       "Bash(list *)",
       "Bash(tree *)",
       "Bash(sed -n *)",
+
+      "Read(//var/mnt/data/projects/settled-reach/**)",
+      "Read(//tmp/**)",
+      "Grep",
+      "Glob",
+
       "Skill(git-commit)",
       "Skill(whats-next)",
       "Skill(pr-process)",

docs/workshops/system-economic-specialization/system-economic-specialization-workshop-brief.md

@@ -0,0 +1,117 @@
+# Workshop — System economic-specialization field (lore-authored dominant commodity)
+
+*Short, focused workshop. Goal: design ONE authored per-system field (maybe two,
+or a small identity set) that drives each settlement's dominant commodity — so the
+built world stays in step with GTTR/wiki lore instead of being guessed by a
+heuristic.*
+
+**Participants:** Miri, Burnelli-Sheldon, Paula
+
+## Why we're here (the problem)
+
+Layer 2 (#982) needs each settlement's **dominant commodity**, which drives its
+built form (D-233: `BulkClass` × `ProductionUbiquity` → coverage, vocabulary,
+spread). We tried deriving it from `economic_role` + corp-HQ + `planet_class`.
+That heuristic **can contradict authored lore**: it could turn the core's
+breadbasket into a foundry, or make the MVG homeworld about cereal products.
+
+**Decision already taken (the frame — don't relitigate):** authored lore WINS.
+The dominant commodity for a system is **authored**, lore/GTTR-consistent. The
+role+corp+planet_class heuristic survives only as a *fallback / weighting basis*
+for unnamed, generic worlds. This workshop designs the authored layer.
+
+## The shape the human wants (design within this)
+
+- **ONE authored field per system** = the system's economic specialization /
+  dominant-commodity character. Lore-driven, GTTR-consistent.
+- **Maybe a SECOND field — `economy_size`** — a modifier. Decide if it earns its
+  place and what it modifies.
+- **Within a system, bodies vary *weightedly*** around the system field — an
+  industrial system can still have a farming moon, but the system reads as its
+  authored character. The variation is weighted, not free.
+- A **300-system content pass** fills this field (Miri + Paula + Burnelli),
+  staying in step with the wiki/GTTR. Keep the authoring burden to ~1 field
+  (2 max, or a tight identity set per Q5) per system.
+
+## Frozen-amber caveat (pin this)
+
+The field is **static, authored, t=0** — read at build time into `systems.db`,
+never the rolling economy. Derivation must stay deterministic (D-010). No
+`PressureState`, no tâtonnement, no save-state.
+
+## Questions for participants
+
+Each participant answers all five from their lens (Miri = lore/GTTR truth;
+Burnelli-Sheldon = economic credibility + mapping + weighting + integration;
+Paula = faction/political economy). Be concrete — name the field, propose its
+vocabulary, name worlds that must be pinned.
+
+**Miri / Burnelli-Sheldon / Paula:**
+
+1. **Q1 — The field.** What *exactly* is the single per-system field? Its
+   vocabulary (a commodity id? a specialization category? primary ± secondary?),
+   and how it maps to `dominant_bulk_class` + `dominant_production_ubiquity`
+   (D-233). Does it reuse/replace the existing `system_economy.economic_base_primary
+   /secondary` (already authored, free-text) or supersede it? One value or a small
+   set?
+2. **Q2 — Within-system weighted variation.** How do bodies vary around the
+   system field? What weights apply (body `economic_role`, `planet_class`, corp
+   HQ, faction)? How is a minority body (the farming moon in a foundry system)
+   expressed without breaking the system's authored read? Must be deterministic
+   (seed + static data only).
+3. **Q3 — `economy_size`?** Is a second field warranted? If so, what does it
+   modify — coverage density (D-220), count of specialized blocks, labor draw
+   (D-233 residential follow-on) — or is it already covered by `population` /
+   `prosperity_baseline`? Decide in or out; don't add a field that duplicates one.
+4. **Q4 — The content pass.** The authoring workflow for ~300 systems: who
+   authors what (Miri = lore/GTTR truth; Paula = faction/political economy, esp.
+   faction homeworlds & capitals; Burnelli = economic credibility + the
+   field→commodity mapping + the weighting model); how to stay GTTR-consistent;
+   where the source lives (wiki TOML → `import_economics.py` → `systems.db`
+   column); CI guardrails (every system has the field; values in-vocabulary;
+   flag systems where the field contradicts existing authored prose).
+5. **Q5 — Do we want a couple MORE identity fields while the pass is open?**
+   We're touching all ~300 systems by hand once. What other 1–2 authored
+   per-system fields would meaningfully flesh out **identity** and are cheap to
+   fill in the same pass — most obviously the **architecture / cultural
+   character** (the D-232 trait-template / `atlas_body_trait_bias` "hero" pin,
+   so a system reads architecturally as canon, not heuristic), and possibly a
+   dominant cultural/political register. Propose the *minimal* high-value set,
+   with the same system-level + within-system-weighted-variation shape. **Guard
+   against scope creep:** every field is authoring burden ×300 and a maintenance
+   surface — only fields that genuinely change how a system reads earn a slot,
+   and each must degrade to a heuristic/default for unauthored systems. Decide
+   the field set; don't gold-plate.
+
+## Build on what exists (don't restart)
+
+- **#982 baseline (provisional, uncommitted):** the role+corp+planet_class
+  heuristic in `import_economics.py` + `city_context_reader.rs` — becomes the
+  fallback/weighting basis, not the primary driver.
+- **D-233** (economic signal → block-fill; the 2026-05-31 amendment defines the
+  current heuristic — this workshop will re-amend it).
+- **D-184** (commodity catalogue), **D-199** (economic read-set), **D-200**
+  (build-time tier-1 denormalization), **D-176/D-178** (rolling economy — OUT of
+  scope here, runtime only), **D-220** (density), **D-197** (prosperity),
+  **D-232** (architecture-flavor / trait templates — relevant to Q5).
+- `system_economy.economic_base_primary/secondary`, `bodies.economic_role`,
+  `bodies.planet_class`, `corp_presence.primary_operation` — the static signals
+  that already exist (Burnelli's earlier data memo has the full inventory).
+
+## Workshop Format
+
+Quick, two rounds max, with a mandatory user checkpoint between.
+
+- **Round 1 — Opening positions.** Each participant reads this brief + the
+  decisions/data for their angle and writes their answer to Q1–Q5 to
+  `docs/workshops/system-economic-specialization/{agent}-round1.md`. Concrete:
+  name the field, its vocabulary, the weighting model, the Q5 field set, and the
+  worlds that must be pinned. Qatux documents; the lead presents findings to the
+  human before any Round 2.
+- **Round 2 — Synthesis / convergence (only if the human approves).** Resolve
+  forks surfaced in Round 1 into a single proposed field schema + content-pass
+  plan. Output → a D-record (re-amending D-233 + any new record) and the ticket
+  set (SI).
+
+Qatux (documenter) and SI (ticket prep) join and are never dismissed early.
+Sonnet-pinned participants.