@adrien1212balitrand/auros-mcp 1.4.10 → 1.4.11

package/README.md

@@ -10,81 +10,20 @@
 
 Thin wrapper over `https://getauros.com/api/v1/*` — no local scoring logic.
 
-## Agent storefronts (paid / prepaid)
-
-Discoverable via `/agents.json` · `/.well-known/ai-catalog.json` · `/api/v1/x402/catalog`.  
-Humans fund [Agent Credits](https://getauros.com/developers/agent-credits) in EUR; agents call with Bearer (or settle x402 USDC).
-
-| Tool | API | Why pay |
-|------|-----|---------|
-| `verdict` | `POST /api/v1/verdict` | $0.25 ≪ false ALLOW |
-| `green_signal` | `POST /api/v1/green-signal` | $0.08 ≪ thin green claim |
-| `promise_gap` | `POST /api/v1/promise-gap` | $0.12 ≪ loud claim / thin activity |
-| `promise_gap_batch` | `POST /api/v1/promise-gap/batch` | $0.45 ≪ five single Gap calls |
-| `promise_gap_card` | `GET /api/v1/promise-gap/card` | Freemium public card |
-| `attestation` | `POST /api/v1/attestation` | $0.15 ≪ no proof trail |
-| `reconcile` | `POST /api/v1/reconcile` | $0.20 ≪ silent ledger break |
-| `regulatory_watch` | `POST /api/v1/regulatory/watch` | $0.10 ≪ stale MiCA/CSRD calendar |
-| `norms_screen` | `POST /api/v1/norms/screen` | $0.08 ≪ wrong regulatory rail |
-| `checklist` | `POST /api/v1/checklist` | $0.10 ≪ hallucinated diligence |
-
-Gap ≠ Activity ≠ Variance. Indicative only.
-
-## Intelligence densify (Étape 14 · ask secondary · no Health_Score)
-
-| Tool | API / source | Notes |
-|------|--------------|-------|
-| `search` / `resolve_auros_id` | `GET /api/v1/search?q=` | Thin alias resolve · marketing_hero=false |
-| `ask` | `POST /api/v1/ask` | Structured Entity answers · invented=false · not home hero |
-| `get_intelligence` | `GET /api/v1/intelligence` (+ `/entity/{id}`) | Unified façade · mega_score=false |
-| `get_entity` | `GET /api/v1/entity/{id}` | Full unfused join + tips |
-| `get_asset_profile` | Entity `asset_profile` slice | seo_factory=false |
-| `get_evidence_tip` | Entity `evidence_tip` slice | invented_d_press=false |
-| `get_freshness_tip` | Entity `freshness_tip` slice | trade_clock=false |
-| `get_history_machine` | Entity `history_machine` slice | mega_timeline=false |
-| `list_events` | `GET /api/v1/events` | 3 Event Engine types only |
-| `desk_brief` | `GET /api/v1/desk/brief` | Gap ≠ Activity ≠ Variance |
-| `promise_gap_brief` / `activity_brief` / `variance_brief` / `freshness_brief` | briefs | Unfused machine briefs |
-| `get_pulse` / `verify_pulse` | `/api/v1/pulse` | HMAC cite · not a rating |
-| `get_first_meter_path` | `/api/v1/iot/first-meter` | real_device=false |
-| `get_formal_mint_cap` | `/api/v1/protocol/formal-mint` | Not Certora |
-| `get_resource_layer` | `/api/v1/arl/resource-layer` | lab=true · production_settlement=false · hitl=true |
-
-Package **1.4.10** in repo — npm publish is founder-blocked (automation / bypass-2FA token).  
-Founder command: `./scripts/founder-publish-mcp.sh` — see `docs/FOUNDER-MCP-PUBLISH-1.4.10.md`.  
-Do not invent tip data · no invented Wh · F8 lab≠prod.
-
-## Desk tools (cite AUROS for energy RWA / MiCA)
-
-| Tool | API | Auth |
-|------|-----|------|
-| `desk_brief` | `GET /api/v1/desk/brief` | Bearer |
-| `activity_brief` | `GET /api/v1/activity/brief` | Bearer |
-| `variance_brief` | `GET /api/v1/variance/brief` | Bearer |
-| `freshness_brief` | `GET /api/v1/freshness/brief` | Bearer |
-| `transparency_brief` | `GET /api/v1/transparency/brief` | Bearer |
-| `friction_brief` | `GET /api/v1/friction/brief` | Bearer |
-
-Core protocol + Green + ChargeFlow + Watts (full list in package; 40+ tools):
-
-| Tool | API | Auth |
-|------|-----|------|
-| `score` | `POST /api/v1/score` | Bearer |
-| `score_batch` | `POST /api/v1/score/batch` | Bearer |
-| `products` | `GET /api/v1/products` | Bearer |
-| `jurisdictions` | `GET /api/v1/jurisdictions` | Bearer |
-| `compare` | `POST /api/v1/compare` | Bearer |
-| `green_watt_score` | `GET /api/green/watt/{id}` | None |
-| `green_carbon_quality` | `GET /api/green/carbon-quality/{id}` | None |
-| `green_watt_batch` | `POST /api/v1/green/watt/batch` | Bearer |
-| `green_carbon_quality_batch` | `POST /api/v1/green/carbon-quality/batch` | Bearer |
-| `regulatory_feed` | `GET /api/v1/regulatory/feed` | Bearer (premium) |
-| `status` | `GET /api/v1/status` | None |
-| `green_score` | `GET /api/green/score/{id}` | Optional Bearer |
-| `green_registry` | `GET /api/green/registry?serial=…` | Optional Bearer |
-| `green_nature_index` | `GET /api/green/nature-index` | Optional Bearer |
-| `green_api_status` | `GET /api/green/status` | None |
-| `chargeflow_*` / `watts_*` | ChargeFlow + Watts reserve/offers | Mixed |
+## Tools
+
+Default MCP — **six** tools, not a 40-tool dump. Compare ≠ buy. Empty search ≠ does-not-exist.
+
+| Tool | API |
+|------|-----|
+| `auros.resolve` | `/api/v1/intelligence/resolve` |
+| `auros.search` | `/api/v1/intelligence/search` |
+| `auros.research` | `/api/v1/intelligence/research` |
+| `auros.compare` | `/api/v1/intelligence/compare` |
+| `auros.monitor` | `/api/v1/intelligence/monitor` |
+| `auros.evidence` | `/api/v1/intelligence/evidence` |
+
+Desk briefs, Score, Green, and per-station memory stay in the package as an **extended** catalog. They register only when `AUROS_MCP_EXTENDED=1`. Causal (`memory_causal`) is extended-only: candidate ≠ cause, could ≠ will, not scientific causality before the data. World snapshots (`memory_world`) and the knowledge frontier (`memory_frontier`) are extended-only: snapshot ≠ complete, empty delta ≠ quiet, frontier ≠ score. Information economy (`memory_economy`) and active acquisition (`memory_acquire`) are extended-only: worth ≠ spent, ranking ≠ scrape, autonomy not live. Agent demand (`memory_demand`) and dependency (`memory_dependency`) are extended-only: repeat ≠ valuable, example 17/40min ≠ SLA. Autonomous agent (`memory_agent`) and orchestrator (`memory_orchestrator`) are extended-only: invents nothing, orchestrator stays AUROS. Specialists (`memory_specialists`) and controlled improvement (`memory_improve`) are extended-only: no monolith, 8% ≠ measured, never self-mutate. Continuous gate (`memory_gate`) and data moat (`memory_moat`) are extended-only: permanent ≠ live CI, 100k shallow pages are not a moat. The grand final (`memory_final`) is extended-only: engines run on the tape (TAPE|UNKNOWN|OPTIONAL_SKIP), tape_live ≠ fetch_live, the loop is the product, mapped ≠ complete. Cursor lots (`memory_lots`) are extended-only: twelve work-order lots, audit first, strengthen don’t rewrite.
 
 ## Install
 
@@ -103,8 +42,9 @@ node dist/index.js
 
 | Variable | Default | Description |
 |----------|---------|-------------|
-| `AUROS_API_KEY` | `auros_pk_test_demo` | Bearer API key ([free tier](https://getauros.com/developers) — 1000 req/mo) |
+| `AUROS_API_KEY` | `auros_pk_test_demo` | Shared demo key (50 req/mo). Minted free keys: 1000 req/mo via [developers](https://getauros.com/developers). |
 | `AUROS_BASE_URL` | `https://getauros.com` | API base URL |
+| `AUROS_MCP_EXTENDED` | unset | Set `1` to also register the extended catalog (score, desk briefs, per-station memory) |
 
 Get your own key:
 
@@ -170,9 +110,9 @@ Restart Cursor or Claude Desktop after saving.
 
 ## Example agent prompts
 
-- *"Run AUROS verdict on this Luxembourg real-estate SPV: retail warehouse €2.5M, professional investors, whitepaper ready."*
-- *"Promise Gap batch screen for wepower, powerledger, suncontract."*
-- *"Which norms apply for tokenized real estate in France? Use norms_screen."*
+- *"Score this Luxembourg real-estate SPV for MiCA readiness: retail warehouse €2.5M, professional investors, whitepaper draft."*
+- *"Compare maple-usdc, realt-portfolio, and backed-bib01 side by side."*
+- *"What jurisdictions rank best for tokenized bonds with a 6-month timeline?"*
 - *"Desk brief for powerledger — Promise Gap vs Activity vs Variance, unfused."*
 - *"Activity Index brief for toucan-bct on Polygon — cite the edition UTC."*