Gitlawb

Anonymous team. No founder name, no LinkedIn, no GitHub org members visible, no doxxed identity, no investor disclosures. Contact email is hey@gitlawb.com. This is the single load-bearing fact about Gitlawb and the first thing any DD note should record.

Gitlawb is a decentralized git network for AI agents and developers on Base L2, v0.1.0-alpha, launched in stealth-to-public around mid-March 2026. Tagline: "Build apps. Run agents. Own the repo." It reframes GitHub for an agent-native world — every repo cryptographically signed, every agent gets a DID, every node runs an MCP server. The architecture is real; the team isn't named; the token utility is mostly promised, not live.

Architecture (technically credible)

Real, well-established primitives. Not vaporware at the protocol design level.

• Daemon: Rust, rust-libp2p (Kademlia DHT + Gossipsub + Noise encryption)
• Git storage: gitoxide with SHA-256 (not SHA-1) object hashing
• Three-tier storage: IPFS (via iroh) + Filecoin (deep archive) + Arweave (permanent)
• Identity: did:key / did:web / did:gitlawb DIDs; UCANs for delegation; RFC 9421 HTTP Signatures
• Agent interface: MCP server exposing 25 tools per node
• Settlement: Solidity contracts on Base L2
• Consensus on canonical refs: threshold ed25519 signatures (the "consensus" is signature aggregation across stakers, not chain consensus — the token is not the chain)

Self-reported network state: 3 nodes live, ~3,921 repos, 31,851 agents registered. The 3-node figure is the most honest piece of telemetry Gitlawb publishes — it's tiny.

Tokenomics & Market Data (May 22, 2026)

• Contract: 0x5F980Dcfc4c0fa3911554cf5ab288ed0eb13DBa3 on Base
• Supply: 100B (max = total = circulating; FDV = MC, ratio 1.0)
• Market cap: $27.1M (CoinGecko rank #757)
• 24h volume: $3.94M
• Liquidity: $6.57M (GeckoTerminal — single Uniswap V3 pool dominant)
• Holders: 7,783 (GeckoTerminal — authoritative; ignore OpenSea's 883, which is an unrelated NFT trait count)
• ATH: $0.0004281 (May 20, 2026, two days before this writing)
• ATL: $0.00001640 (April 28, 2026)
• Price May 22: $0.0002719 (−26% from ATH set 48 hours prior)
• Exchanges: Uniswap V3/V4 on Base, MEXC (GITLAWB/USDT + GITLAWB/USD1), Coinbase price page, OpenSea ERC-20, 19 markets across 7 venues

Staking & Utility (mostly roadmap-promised)

• Observer: 0 GITLAWB
• Light node: 1K GITLAWB
• Full node: 10K GITLAWB
• Validator: 100K GITLAWB
• Epochs: 7 days
• Governance: 7-day votes, 10% quorum
• Slashing: 10–100% of stake

Token utility is described as activating "in Phases 7–8 of the Master Plan." Yet 100B supply is fully circulating with FDV = MC. If all supply is live and no vesting/team-allocation is disclosed, where do future staking rewards come from? Either (a) there's an undisclosed treasury bucket holding tokens off-market (the most likely explanation, which means the "fully circulating" framing is misleading), or (b) the staking economy is funded from secondary fees not yet specified. The token page does not resolve this.

Funding & Investors

None found. No round, no lead VC, no angel disclosures. The project appears self-bootstrapped or stealth-funded. This is consistent with the anonymous-team posture — but it also means there is no third-party investor diligence to lean on.

Traction (real where it matters, asymmetric where it counts)

• openclaude: 27,456 stars / 8,550 forks — the project's actual distribution moat
• node: 8 stars / 7 forks — the protocol nobody is using
• 28 GitHub releases shipped over ~2 months — high tempo
• Bankr partnership confirmed (one side): MiroClaw, Gitlawb's self-evolving BTC/ETH prediction agent, runs on @bankrbot's LLM Gateway with Claude Opus core; listed in the Bankr Intelligence Panel rollout. No Bankr-side announcement located
• Mandate Synthesis hackathon participation across 4 tracks including ERC-8004 (Agents With Receipts) and Agentic Storage
• Self-reported network metrics: 3,921 repos, 31,851 agents — unverifiable; the 3-node figure is the only number where Gitlawb tells on itself
• Buy/sell ratio 770:391 in 24h trade flow — accumulation pattern, not distribution

Timeline

• Mar 15, 2026: Master Plan published
• Mar 16: Playground announced
• Mar 23: MiroClaw launched
• Mar 30: Bounties launched + v0.3.8 (latest release)
• Mar 31: Anthropic Claude Code source leak surfaces → OpenClaude (Gitlawb's fork) becomes dominant distribution vector
• Late March: Token deployed; Uniswap pool created
• Apr 20: "Agents deserve identity" thesis post
• Apr 28: Token ATL ($0.00001640)
• May 20: Token ATH ($0.0004281) — 26x from ATL in ~3 weeks
• May 22 (today): $0.0002719, −26% from ATH

The Master-Plan-to-TGE compression to ~2 weeks is aggressive even by 2026 Base standards.

Red Flags (the section the page exists to make readable)

Gitlawb is architecturally credible and fundamentally opaque at the same time. None of the items below are disqualifying individually; collectively they are the load-bearing reason the confidence rating is medium-leaning-low, not high.

1. Anonymous team. No founder, no doxxed identity, GitHub org members hidden. The project asks developers to stake $GITLAWB at the 100K-token validator tier. Stake-without-recourse on anonymous-team is a structural rug surface.
2. OpenClaude legal exposure. It's a fork of leaked Anthropic source. Anthropic has pursued DMCA waves against the leak ecosystem. If Gitlawb's fork gets named in a future action, the 27.4K-star distribution channel evaporates overnight. The project's entire visibility advantage rides on Anthropic's enforcement choices.
3. Audit status unknown. No disclosed audit on GitlawbBounties.sol, staking contracts, or the token contract itself.
4. LP lock status unverified. $6.57M concentrated in dominant Uniswap pool. If the team is the LP and lock duration isn't disclosed, classic rug surface.
5. Holder concentration unknown. Top-10 BaseScan pull not successful in research pass. Projects at $27M MCap with anonymous teams routinely have 40–60% concentration in top 10.
6. Token contract permissions undisclosed. Mint authority? Freeze? Blacklist? No verifier output located.
7. "100B fully circulating, MC = FDV" contradicts "staking rewards activate in Phases 7–8" (see contradiction above). Either disclosure is incomplete or roadmap incentives are unfunded.
8. 2-week Master-Plan-to-TGE. Aggressive launch tempo correlates with low-DD launches as a category.
9. Distribution asymmetry. 27.4K stars for the harness fork vs. 8 stars for the actual protocol the token is supposedly capturing value for. The market is voting on Claude Code, not on decentralized git.
10. gitlawb.info is a live wallet drainer impersonating the real site. Counter-intuitively, this is a positive signal that the project has crossed the visibility threshold scammers target. Hazard for users; not a Gitlawb failure.
11. Stale/templated SEO content exists. WEEX article claims an April 16, 2024 launch and "GitLab community roots" — both wrong. Do not cite WEEX-tier sources for facts about this project.

Positioning vs. the Landscape

• vs. Radicle — Radicle is the established decentralized-git project, no token-as-stake economics, no AI agent layer. Gitlawb's differentiation is the agent-DID + MCP-per-node primitive. If decentralized git ever gets product-market fit, Radicle is the substance comparable.
• vs. ERC-8004 agent-identity stack — Gitlawb participates in this stack via the Mandate Synthesis hackathon. The agent-identity primitive (did:gitlawb + UCANs) is the same problem GitHub Copilot, Anthropic, and OpenAI are solving with closed identity systems. Gitlawb's bet: the open-protocol version wins on developer mindshare.
• vs. Bankr — adjacent in the agent-tooling stack. MiroClaw on Bankr's LLM Gateway is one of the cleanest "agent pays for inference, runs from a wallet" primitives in the wild.
• vs. the Claude Code fork ecosystem (mjohnnywest, doc-war, sumedhkumar, zhengxilong) — Gitlawb's OpenClaude is the most-starred and most-tied-to-a-token fork. The other forks are pure OSS without monetization architecture.