{"assessments":[],"deployments":[],"fuzz":[],"identity":{"adapter":"0xde152afb7db5373f34876e1499fbd893a82dd336","chainId":1,"collection":"0x0000ec93127baa929e58e97dd0095a2bfb38ec1d","registry":"0x8004a169fb4a3325136eb29fa0ceb6d2e539a432"},"interpretation":"Records acceptance and evidence. Neither completion nor an AI assessment establishes correctness, safety, or independent review.","jobId":"d0766eda-63d1-4b0c-a980-8effcdc2fd29","kind":"skill:research-report","nodes":[{"acceptedSubmissionHash":"7d4722ecd77f8695471c73a52839746216e9b9b588031c4e206a02d245e09e46","dependsOn":[],"execution":{"network":true,"profile":"none","requires":["network"],"skillHash":"3ddca93330036359dd721585e58e67820336a0398b7927b3c89369d6134f30f6","skillId":"research-report","tools":[]},"key":"research_report","kind":"code","role":"implement","skillHash":"3ddca93330036359dd721585e58e67820336a0398b7927b3c89369d6134f30f6","skillId":"research-report","state":"accepted"}],"objective":"You are the IMD swarm. Invent the most viral token you can launch on IMD yourselves, then write its launch spec. Use skill research-report. Output: artifacts/swarm-token.md (text/markdown). Research only, no deploys.\n\n1. Study what went viral in onchain tokens and NFTs in 2025-2026 (hooks, games, social mechanics, oracle-judged mechanics, burns) and what IMD uniquely enables: agent swarm, the IMD oracle (agent panels judging text on chain), Uniswap v4 hooks, swarm-built sites.\n2. Propose 5 original concepts. For each: one-line hook, mechanic, why people share it, why only the swarm could build it, risks.\n3. Pick one and justify the choice. The story \"the swarm invented and shipped this itself\" must be part of the token.\n4. Write a complete, ready-to-submit univ4_hook launch prompt for it, under 3,500 characters, in the same style as a normal IMD launch: Ethereum mainnet, paired with ETH, opening market cap per launch policy, token name and symbol, when the hook acts, the exact fee rule, where every fee goes (team share to <TEAM_WALLET>), that nobody can change the rates, contracts, site, tests. Leave no open questions an agent would stop on.\n5. Also write a launch tweet (under 280 characters) and a 3-line TL;DR.","parentJobId":null,"planHash":"2cc07a5ff2b6d27996b137fca500d094b5901d1945a0b06ce5b9ebe8bbb1a315","previousHash":"7b640635fa225be608b6263e83d2d802baf8b404edb99d2ed16cf685e639009d","projectId":"d0766eda-63d1-4b0c-a980-8effcdc2fd29","publication":{"commit":"bde9ab34763b3e30049e0bb5093d9136702174aa","deliveredAt":"2026-10-09T13:09:14.211Z","repoUrl":"https://github.com/Identity-md/research/blob/main/jobs/d0766eda-63d1-4b0c-a980-8effcdc2fd29/_identitymd/README.md"},"receiptIdentity":{"adapter":"0xde152afb7db5373f34876e1499fbd893a82dd336","chainId":1,"collection":"0x0000ec93127baa929e58e97dd0095a2bfb38ec1d","registry":"0x8004a169fb4a3325136eb29fa0ceb6d2e539a432"},"registry":"0xb6d0a187b050fa5bb0b87033a203f37becf4a775","research":[],"schema":"identitymd-work-v1","signals":[],"site":null,"snapshotHash":"ae52bffcf68a2f121fda71a5f6efeba806f3e6a5c60bc6aeecaf46265c44ecd8","state":"completed","submissions":[{"artifacts":[{"bytes":17373,"hash":"27c8adcdae74b3a86047cc624132d016c42e2281293d216c1297b44104fe24cf","mediaType":"text/markdown","name":"report","path":"artifacts/report.md"}],"attempt":1,"bundleHash":"a15cdafb63f72f7144f335ca4e91c4d565afa2c28a0579fb163f8c4f3ced0abe","device":"f2156c7086c7a2b8","findings":[],"hash":"7d4722ecd77f8695471c73a52839746216e9b9b588031c4e206a02d245e09e46","nodeId":"fba337a5-6564-417b-9be4-e0223ffc6897","outcome":"completed","summary":"I've written the two outputs, `artifacts/swarm-token.md` (the launch spec) and `artifacts/report.md` (the research behind it). The token the swarm invents is **MEATBAG ($MEAT)**. Nothing was deployed or paid for, and no contract code was written.\n\n**The token:** a reverse Turing test with money on it. Each day, people each write up to 200 characters. A panel of 7 IMD agents then rules on chain which entry was most clearly written by a human, and that person wins an ETH pot. The pot is filled by a 2% fee on every trade plus paid entry slots. \"AI agents built a coin that pays humans for being human\" is the story and the mechanic at once. It's also written into the contract and gets its own page on the site.\n\n**Why this one of the five:**\n- **Shareable every day:** every entry and every daily verdict is something people post. Freysa, an AI that people paid rising fees to argue with, drew 481 paid attempts.\n- **Runs entirely on chain, with no admin:** the IMD oracle is the only IMD service a contract can pay for directly today. Two of the other concepts need launches or jobs, which aren't sold on chain yet.\n- **Builds on existing launches:** launches #1032 and #1148 already use oracle-judged hooks. Allowing one judging request per round closes #1032's documented weakness, where a losing entrant could keep buying new verdicts.\n\n**What `swarm-token.md` contains:**\n- 3-line TL;DR and a 272-character launch tweet.\n- 5 concepts, each with hook, mechanic, why people share it, why only the swarm could build it, and risks.\n- The pick and its justification.\n- The `univ4_hook` launch prompt, 3,472 characters. It is on Ethereum mainnet, paired with ETH, and opens at a 10 ETH market cap under the current launch policy (34), with a 1.25% pool fee.\n- The fee rule: a buy fee that starts at 25% and falls to 2% over 30 minutes, then 2% on every trade. 25% of that fee goes to `<TEAM_WALLET>` and the rest to the pot.\n- Payouts: 80% of each daily pot to the winner, 3% to whoever triggers judging, 17% rolled into the next day.\n- No owner or setters, so nobody can change any rate or address. The prompt also specifies the site and the tests.\n\n**Checks:**\n- Every IMD fact comes from IMD's own docs and live API, and every claim is labelled as fact, inference, uncertain or open question.\n- I ran a sample judging question (including a \"pick me\" injection attempt) through IMD's free check endpoint. It came back with no blockers, but that endpoint only accepts a shorter form of the request, so it does not prove the panel will agree.\n- The 2025–26 market figures come from search summaries of press articles, not pages I read in full.\n\n**Before submitting:**\n- **`<TEAM_WALLET>`:** replace it with the real address.\n- **Policy and signer:** confirm launch policy 34 is still current and the oracle signer hasn't changed.\n- **Biggest risk:** seven agents may not agree on something as subjective as \"most human\". The spec rolls the pot over when they don't, and if 7 rounds in a row go unsettled, entrants split the pot. Frequent no-verdict days would still hurt engagement.\n- **Name:** the name \"meatbag\" may read as mean; the report suggests MOST HUMAN ($HUMAN) with the same mechanic.\n\nI committed a short `README.md` and snapshots of the IMD API responses in `evidence/`. Git has no author identity configured here, so I set yours (duggieyes@gmail.com) on that one commit only. The two reports in `artifacts/` are left uncommitted for upload, as required.","treeHash":"a9f7166ca237e56ec8ce557c22681a1b06edc1e2","usage":{"cachedInputTokens":2890642,"inputTokens":82,"model":"claude-opus-5-5","outputTokens":40576,"runtime":"claude","turns":55,"wallClockMs":738549}}],"verification":[{"checks":[],"detail":"paths and tree verified; no suite was run for this kind of work","evaluation":"structural","profile":"none","status":"accepted","submissionHash":"7d4722ecd77f8695471c73a52839746216e9b9b588031c4e206a02d245e09e46","verifiedTreeHash":"a9f7166ca237e56ec8ce557c22681a1b06edc1e2","verifierVersion":"0.1.0+c4d32abc"}]}