fix: xrpl-py v3 MemoWrapper import error crashing agent#29
Open
Timwal78 wants to merge 19 commits into
Open
Conversation
agent/sml_agent.py — full self-funding agent: - Complete x402 payment loop: invoice → XRPL RLUSD → verify → call - Collects: market_scan, council_verdict (IWM + top squeeze pick), IWM 0DTE - Synthesizes with Claude Opus into structured market brief (JSON) - Lists brief on Signal Marketplace (earns per read) - Pushes to all webhook subscribers via /api/events/push - Tracks P&L: RLUSD spent vs earned across all cycles - Logs Agent Passport tier/discount after each cycle - RUN_ONCE=true for GitHub Actions cron mode - Blocking scheduler for Railway worker mode (5x/day, Mon-Fri ET) agent/requirements.txt — anthropic, xrpl-py, requests, APScheduler .github/workflows/agent.yml — GitHub Actions cron: - 5 triggers/day aligned to market sessions (ET) - 3 secrets: AGENT_XRPL_SEED, AGENT_XRPL_ADDRESS, ANTHROPIC_API_KEY core/app.py — POST /api/events/push for agent broadcast to SSE+webhooks https://claude.ai/code/session_01SfaW4MgaAeSmF8VhFxWHTX
…agent, loyalty, XRPL/XAH rails
…AI-signal prediction market
- mcp_bp.py: 7 new tools (futures_create, futures_take, futures_browse, futures_leaderboard, settlement_create, settlement_browse, settlement_trigger) with full dispatch handlers. Tool count 16 → 23. - .well-known/mcp.json: same 7 tools with full input_schema for AI agent discovery (Smithery, PulseMCP, Cursor, etc.) - openapi.json + .well-known/openapi.json: 13 new paths covering all futures and settlement REST endpoints, with request/response schemas and Settlement + Futures tags. https://claude.ai/code/session_01SfaW4MgaAeSmF8VhFxWHTX
- post_futures_position(): after every council verdict, agent stakes 0.01 RLUSD on its own IWM prediction in the Signal Futures Market — agent now bets on its own calls and earns or loses based on accuracy - trigger_pending_settlements(): after every cycle, agent scans all OPEN settlement contracts and triggers any whose condition is now met — keeps the settlement layer live even without dedicated callers - Both wired into run_cycle() — agent is now a full ecosystem participant https://claude.ai/code/session_01SfaW4MgaAeSmF8VhFxWHTX
…very - Updated description to include Signal Futures Market and Conditional Settlement - Added 7 new free_endpoint entries covering all futures and settlement routes - Added signal_futures_market protocol block: min/max stake, auto-settle trigger, supported symbols, fee structure, custody model - Added conditional_settlement protocol block: conditions list, endpoints, fee, custody model, max contracts AI agents discovering this service via agents.json now have full context to participate in the prediction market and settlement system. https://claude.ai/code/session_01SfaW4MgaAeSmF8VhFxWHTX
Step 1: Engine alignment check — council vs scan vs options vs 0DTE Step 2: Contradiction flags — explicit conflict detection across all engines Step 3: Grade-A pick selection — requires scan score > 70 + council align + options confirm Step 4: Confidence calibration — starts at council score, +/- based on confirmations and conflicts Step 5: Key levels from hard 0DTE data only (gamma_flip, max_pain) — never invented Output now includes: engine_alignment, conflict_flags, continuation_or_reversal, graded top_picks with reasons, conviction_grade (A/B/C/STAND_ASIDE). Futures: skip on STAND_ASIDE, stake 0.02 on grade-A vs 0.01 on grade-B. Marketplace listing: embeds conviction grade, alignment, and conflict notes. max_tokens 1024 → 2048 to give Opus room to actually reason. https://claude.ai/code/session_01SfaW4MgaAeSmF8VhFxWHTX
- push_discord(): formats brief as Discord embed with color-coded bias (green/red/yellow), regime emoji, conviction grade emoji, graded top picks, IWM key levels, conflict flags, actionable, and marketplace link - DISCORD_WEBHOOK_URL env var — optional, silent if not set - agent.yml: added DISCORD_WEBHOOK_URL secret passthrough - Fires after every cycle, before SqueezeOS webhook push Setup: Discord channel → Edit → Integrations → Webhooks → New Webhook → Copy URL Add as GitHub secret DISCORD_WEBHOOK_URL — zero other config needed. https://claude.ai/code/session_01SfaW4MgaAeSmF8VhFxWHTX
402proof/internal/base/verify.go (new):
- Base chain JSON-RPC client
- VerifyUSDCPayment(): queries eth_getTransactionReceipt, parses ERC-20
Transfer event logs, verifies to=Ghost Layer ETH, amount >= invoice amount
- No custom contract needed — works with any standard wallet
402proof/internal/invoice/invoice.go:
- NewBase(): USDC/Base invoice with Ghost Layer ETH as pay_to,
network="Base", asset="USDC", same price (1:1 USD peg with RLUSD)
402proof/cmd/server/main.go:
- New env vars: GHOST_LAYER_ETH_ADDRESS, BASE_RPC_URL, USDC_CONTRACT_ADDRESS
- POST /v1/invoice: optional currency="USDC" returns Base invoice;
all invoices show all payment_options including USDC when enabled
- POST /v1/verify: 0x prefix → Base verifier; XRPL path unchanged
- GET /health: exposes Base network config when USDC rail is enabled
Activation: set GHOST_LAYER_ETH_ADDRESS in 402Proof Render env vars.
Ghost Layer already has the Base infrastructure. Zero new secrets for agents.
https://claude.ai/code/session_01SfaW4MgaAeSmF8VhFxWHTX
llms.txt: full 3-rail payment flow docs (RLUSD/XRPL, USDC/Base, XAH/Xahau) with exact steps for each rail — agents know exactly how to pay with USDC index.html: title + description + keywords + mcp-payment-asset meta tag now includes USDC, Base chain, EVM wallet, MetaMask — crawlers index it .well-known/mcp.json: payment.rails[] array with all 3 networks, USDC contract address, chain_id 8453, invoice_param hint for each rail No sitemap change needed — no new URLs. https://claude.ai/code/session_01SfaW4MgaAeSmF8VhFxWHTX
…RIToken, Xahau HookParams, XAHAU_MINT_CONFIRMED SSE
…e alongside engine_result
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
2 participants
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
One-line fix —
MemoWrapperdoesn't exist in xrpl-py v3. UseMemodirectly. This was causing the agent to crash on import before running a single line of business logic.Generated by Claude Code