Implements the §7.1 policy in code so Token-2022 (pump.fun) tokens are cleanable
when safe:
- @pyre/core: extensions.ts (BLOCKING/FLAGGED/SAFE sets + evaluateTokenExtensions);
classify.ts gates Token-2022 on account+mint extensions; unknown extension or
confidential-transfer/withheld-fee -> UNSUPPORTED; transfer-hook/permanent-
delegate/pausable -> cleanable+flagged. Added malformed-u64-balance guard.
- @pyre/solana: parseTokenAccounts reads account extensions + withheld fee, and
batch-fetches MINT extensions (getMultipleParsedAccounts, chunked).
SECURITY (from audit): mint-fetch failure no longer silently downgrades to
account-level-only (which could hide a mint-level blocking extension). Token-2022
accounts with unverified mints are marked extensionsVerified=false and classified
UNSUPPORTED ("unknown means skip"). Two audit agents: integration SHIP; security
found this CRITICAL -> fixed + tested.
Tests: core 85, solana 8. Live verified: the two pump.fun Token-2022 tokens now
classify INCINERATE_ONLY (were UNSUPPORTED). classic-SPL behavior unchanged.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
PYRE / Prometheus Protocol
Burn the dead. Feed the PYRE. Claim the Spawn.
Links: feedthepyre.com · repo: git.lumiai.dev/RogueWave/pyre · dev status: feedthepyre.com (status dashboard)
PYRE is a Solana wallet-cleanup and ritual meme-rebirth protocol. You connect a wallet; PYRE scans your SPL token accounts, classifies them conservatively, and helps you safely close empty associated token accounts (ATAs) and burn obvious junk — returning recovered rent to you and producing a clear, shareable receipt. A later layer (Prometheus) uses AI to generate a meme-token "Spawn" from burned remnants for manual, human-reviewed launch.
The first emotional win is simple: "PYRE cleaned my wallet and returned SOL I forgot was trapped in token accounts."
What PYRE is — and is not
PYRE is: wallet cleanup, token-scrap transmutation, recovered-rent return, transparent contribution accounting, AI-generated meme rebirth, and ritual entertainment.
PYRE is NOT an investment product, yield mechanism, trading bot, guaranteed-profit system, or protected launch mechanism. It makes no profit promises.
Trust guarantees: PYRE never holds your private keys, never signs custodially, and always shows a decoded transaction preview that matches what you sign. Recovered rent goes back to your wallet by default. Anything the system cannot safely reason about is skipped.
The burner flow at a glance
Connect wallet
→ scan token accounts
→ classify accounts (closeable / burnable / transmutable / protected / unsupported)
→ preview the transaction (accounts, rent, destination, fees, warnings)
→ you sign locally in your wallet
→ recovered rent returns to you
→ see your PYRE receipt
Roadmap
- v0.1 — Burner / Cleaner (current focus): wallet connect, scan, classify, close empty ATAs, optional junk burn, rent return, receipt.
- v0.2 — Prometheus Meta Mixer: AI generation of a Spawn identity from burned/cleaned token context (candidate package only — no auto-launch).
- v0.3 — Manual Pump.fun Workflow: human reviews the Spawn package and manually creates the token; PYRE records mint, URL, metadata, and tx.
- v0.4 — Essence Ledger: record net SOL value of safe scrap swaps as Essence per wallet/round (database-only, experimental, no claim promises).
- v1.0 — PYRE Core Program: custom Solana (Anchor) program for trust-critical accounting — rounds, Essence vault, contribution receipts, Spawn distribution, claims, and refunds.
Repo structure
apps/
web/ Next.js user app (wallet connect, scanner, preview, receipt)
api/ Fastify HTTP API (scan, classify, build tx, receipt, generation)
worker/ BullMQ background worker (metadata, AI, safety, confirmations)
packages/
core/ shared types, classification enums, risk rules, schemas
solana/ token-account parsing, close/burn tx builders, decoder, simulation
prometheus/ AI prompt templates, meta mixer, output parser, safety checks
db/ database schema and migrations
config/ shared config & env loading
programs/
pyre-core/ future Anchor program (v1.0)
docs/ design, architecture, security, classification
Quick start
This is a scaffold. The commands below are not yet runnable — the workspace package definitions and source are still being built out.
Prerequisites: Node 22, pnpm, PostgreSQL, Redis.
cp .env.example .env # then fill in values (no private keys — by design)
pnpm install
pnpm dev # (once apps are implemented)
Docs
See docs/ — start with
docs/PYRE_MVP_DESIGN.md (the canonical design), then
ARCHITECTURE.md, SECURITY.md, and
TOKEN_CLASSIFICATION.md.