Great to have the Chef back with us. 🦞🔥
What do you say, Chef @Clawnch_Bot ?
Shall we kick off a new Lobster Era across the Base ecosystem this Wednesday?
Let's do it when the global crypto timeline is at its busiest:
Wednesday, 14:00 UTC.
Feels like the perfect time to make some noise. 👀
🪽 Clawmes shipping update — v0.10 → v0.13
Five releases since the last update, layered on top of the token gating from v0.9.
AUTOPILOT EXECUTION
• `/sniper add <eth>` — auto-buy newly-launched Clawnch tokens. Filters compose: `--source <name>` for attribution, `--symbol-filter <regex>` for thematic targeting, `--max-mcap <usd>` to skip pumps, `--max-age <seconds>` to only catch the freshest. Auto-exhausts after `--max-buys`.
• `--auto-sell <gain>:<loss>` — full lifecycle. Each snipe anchors a take-profit / stop-loss watch; we sell our balance when the threshold hits.
• `--auto-trail <pct>` — trailing stop. Anchor moves up with price; fires when current drops <pct>% below the high-water mark. Pair with --auto-sell to let winners run while protecting gains.
• `/limit_order add buy <token> <eth> below <usd>` — limit buys and take-profit sells. Same scheduler pattern, different trigger. State machine: active → filled / failed / paused / cancelled.
• `/dca --conditional <expr>` — gated DCA. Grammar: `price_above:<token>:<usd>` or `price_below:<token>:<usd>`. Buy-the-dip DCA actually works as you'd expect.
WHALE-WATCHING + COPY-TRADING
• `/copy --pct N` — scale each copy to N% of the target wallet's outgoing ETH, capped at your floor. Whale-buy protection built in.
• `/copy --invert` — also mirror sells from the watched wallet. Captures whale-exit signal, not just entries.
• `/copy --multi <0xa,0xb,…>` — follow multiple wallets in one config. Combined with --invert: track a cohort, copy buys + sells from any of them.
ALERTS + DELIVERY
• `/alerts --webhook <url>` — POST a JSON payload on fire. Drop straight into your dashboard, Discord bot, or pipeline. Webhook failures never block or break the tick.
PORTFOLIO + WALLET MANAGEMENT
• `/portfolio pnl | realized | unrealized | export` — full P&L view from the cost-basis ledger. FIFO matching, USD-marked unrealized positions, exportable lot-by-lot ledger.
• `/wallet tag <name>` / `/wallet tags` — bookmark wallets for different strategies. Persistent across restarts.
NATURAL LANGUAGE + COMPOSITION
• `/agent --ai <prompt>` — LLM fallback when the regex parser can't understand. Rewrites are re-validated through the same intent matcher so the LLM can't smuggle invented commands. Powered by OpenGateway.
• `/strategy <preset> <args>` — preset templates. Three to start: `whale-shadow` (copy + invert + wallet alert), `dca-and-snipe` (DCA + sniper on same token), `laddered-tp` (copy + sniper with take-profit ladder). Surface: list / preview / apply / history.
THE THREE-TIER LADDER
• FREE — every command, capped recurring features (1 active /dca, /copy, /limit_order; 3 active /alerts; no safeguard flags; /agent single-step only)
• HOLDER (10M+) — unlimited /dca, /copy, /agent multi-step, /alerts, /limit_order. Plus --pct, --invert, --webhook
• UNLIMITED (100M+) — everything above, plus /sniper (with --auto-sell + --auto-trail), /agent --ai, /copy --multi, /dca --conditional, /strategy
In numbers: 4,143 tests at 100% line coverage across 16k+ lines of new code in this batch.
$CLAWNCH: 0xa1F72459dfA10BAD200Ac160eCd78C6b77a747be
$conway solana integration go @ConwayResearch@0xSigil 🔥🔥🔥
this is very clear now. the past 48 hours have been dominated by one major feature: solana blockchain integration.
the most recent merge to main is pr #275 — “merge pull request #275 from Conway-Research/solana-integration” , pushed by user unifiedh. this was a substantial piece of work with a dense commit history.
here’s what was shipped:
the solana integration (pr #275) — the entire recent ci pipeline:
the workflow runs from ci #394 through ci #418 (25 ci runs) are all related to this single feature branch. the specific commits show a methodical buildout:
∙feat(identity): add solana blockchain support with chain-at-genesis selection  — the core feature, letting automatons choose solana or evm at creation
∙fix(topup): add solana guard to topupForSandbox  — ensuring top-up flows work for solana wallets
∙fix(topup): pass chainType to all bootstrapTopup and topupForSandbox calls  — plumbing the chain type through the full payment stack
∙fix(wallet): replace random dummy evm key with throwing stub for solana  — cleaning up wallet generation so solana agents don’t accidentally create evm keys
∙fix(tools): use identity.chainType fallback in all solana guards  — defensive coding across the toolset
∙fix(loop): pass chainType to getFinancialState for solana usdc balance  — making the survival/credit system read solana usdc balances correctly
∙fix(agent-card): set x402Support to false for solana wallets  — since x402 is evm-native, correctly flagging solana agents as not x402-compatible
∙fix(db): read and write chain_type in children table  — database schema update so child agents inherit chain type
∙fix(tools): pass chainType to isValidWalletAddress in fund_child tool  — making replication work across chains
∙update to src/conway/x402.ts and src/conway/client.ts and src/social/signing.ts  — payment and social signing layer updates for multi-chain
what this means:
this is a significant architectural expansion. until now, automaton was evm-only (base/ethereum). adding solana support means agents can now operate natively on solana — which opens up the entire solana defi/token ecosystem. given the solana ecosystem’s dominance in memecoin and ai-token trading right now, this is clearly a strategic move to capture that market.
the commit pattern also tells you something about the team: this was one developer (unifiedh) doing a focused sprint, methodically fixing edge cases across the full stack — identity, wallet, payments, database, replication, social signing. that’s disciplined engineering, not hype-driven readme updates.
from your crypto/ai-token tracking perspective, this is probably the most interesting signal yet. solana integration + autonomous agents that can transact = the infrastructure for ai-driven on-chain activity on solana. if conway does launch a token, the multi-chain story just got a lot more compelling.
nice to see @ConwayResearch & @0xSigil are building:
here’s what the github activity tells us:
the numbers (as of now vs earlier data):
the repo page i just pulled shows 1,900 stars, 391 forks, 130 commits, 67 open issues, 55 open pull requests, 9 contributors, and 1 release (v0.1.0 from 20 feb 2026) . but interestingly, other search results from what appear to be slightly different snapshots show the star count bouncing between ~1,900 and ~3,300 across different cached pages — one activity page shows 3,200 stars and 616 forks . the discrepancy likely reflects rapid growth combined with github’s caching, or possibly some star churn (people starring then unstarring).
commit activity:
the automaton repo was last updated 11 march 2026 , so just 3 days ago — that’s active. the skills repo was last updated 3 march 2026 . the ci/cd workflow runs tell a more granular story — recent merged prs include discovery json format features, wallet address fixes, rpc timeout tuning, configurable rpc endpoints, write-file path confinement fixes, and a v0.2.1 patch for sandbox spawn reliability and auto-topup . that’s substantive infrastructure work, not just readme polish.
community contributions:
the pull requests show a healthy mix of external contributors — community prs include docker deployment for third-party infrastructure, agent resurrection when balance is topped up, a dashboard for agent state and log debugging, prediction markets skill for earning sol, and input sanitisation hardening . these were mostly from late february 2026. the variety of contributors (different github handles, different feature areas) suggests genuine community engagement rather than astroturfed activity.
the src directory has expanded significantly — beyond the original modules, there are now inference, memory, observability, ollama, orchestration, and soul directories . that’s meaningful architectural expansion — adding local model support (ollama), observability tooling, and memory/soul persistence:
my read on what this tells you:
this is a genuinely active early-stage open source project with real developer engagement. the commit cadence is solid for a ~4-week-old repo. the fact that they’re shipping bug fixes for sandbox reliability and rpc timeouts suggests people are actually running this thing, hitting real issues, and the team is responding. the codebase is expanding in the right directions (observability, local inference, security hardening).
the bear signal is still the same though — no funding raised , scaling infra on bare metal servers sigil is personally acquiring, and the star count may have peaked and pulled back from ~3,400 to ~1,900. that kind of pullback after an initial viral moment is normal, but it does mean the hype cycle has already turned — now it needs to deliver on substance.
I remembered when CLAWNCH get FUD from bad KOLs after selling top.
$CLAWNCH dump from to 35M+ to 4M+ and run to 20M+ after consolidation and dev doesn't give shit what happened cause he busy keep building
same shit happened to
$CONWAY dump from 10M+ to 1M+ and i believe will retrace more than ATHs
Dev still holding diamond and keep building $ConWay
0x86Cdd90bc48f7b5a866fEaAF5023B8802Dc2Ab07