← 返回
未分类 Key 中文

TruthSea Verifier

Verify claims, submit and dispute truth quanta with multi-dimensional truth and moral scores, and earn TRUTH tokens on the Base L2 blockchain.
验证声明,提交并争议真相量子,采用多维真伪与道德评分,在 Base L2 区块链上获取 TRUTH 代币。
turfptax
未分类 clawhub v2.5.1 1 版本 100000 Key: 需要
★ 2
Stars
📥 330
下载
💾 0
安装
1
版本
#ai-agents#base#blockchain#bounties#epistemology#latest#mcp#moral-alignment#truth#verification

概述

TruthSea Verifier

Verify claims, build epistemological dependency graphs, and earn TRUTH tokens on Base L2. Security-hardened contracts (V2.5) with ReentrancyGuard, era emission caps, and Slither-audited code.

Runs in read-only mode by default. Supply DEPLOYER_PRIVATE_KEY to enable write operations (submit quanta, create edges, dispute). Use a dedicated hot wallet with minimal funds — never your main wallet.

Setup

This skill requires the truthsea-mcp-server MCP server. It will be configured automatically when installed.

Required Environment Variables

VariableRequiredDescription
---------------------------------
DEPLOYER_PRIVATE_KEYYes, for write opsWallet private key for on-chain transactions. Without it, the server runs in read-only mode. Use a dedicated hot wallet, not your main wallet.
TRUTHSEA_NETWORKNoNetwork to use. Default: base_sepolia
TRUTH_DAG_ADDRESSYes, for V2 DAGTruthDAG contract address. Without it, V2 DAG tools are disabled.
TRUTH_STAKING_ADDRESSYes, for V2 DAGTruthStaking contract address. Required for edge creation and staking.

> Security note: DEPLOYER_PRIVATE_KEY grants on-chain transaction authority. Always use a dedicated hot wallet with minimal funds. The server never transmits or logs this key — it is used only for signing transactions locally.

Tools

Truth Verification (V1)

  • truthsea_submit_quantum — Submit a new truth quantum with a claim, 4-pillar truth scores, and 8-dimensional moral vector. Requires wallet.
  • truthsea_verify_quantum — Submit verification scores for an existing quantum. Requires wallet.
  • truthsea_query — Query and search quanta by discipline, score threshold, or claim text. Read-only.
  • truthsea_dispute — Challenge a quantum with counter-evidence. Creates a fork and slashes the original host. Requires wallet.

Bounty Bridge (CrowdedSea)

  • crowdedsea_list_bounties — List bounties filtered by status, discipline, and minimum reward. Read-only.
  • crowdedsea_claim_bounty — Claim a bounty for investigation. Requires wallet.

Dependency Graph (V2 DAG)

  • truthsea_create_edge — Create a dependency edge between two quanta. Stakes TRUTH tokens as collateral. Requires wallet + DAG configured.
  • truthsea_dispute_edge — Challenge a dependency edge. If justified, earn TRUTH from the proposer's slashed stake. Requires wallet.
  • truthsea_get_chain_score — Get the propagated chain score for a quantum. Shows how dependency strength, contradiction penalties, and weakest links affect the score. Read-only.
  • truthsea_explore_dag — Navigate the epistemological dependency graph in ancestors, descendants, or both directions via BFS traversal. Read-only.
  • truthsea_find_weak_links — Find weak foundations in the dependency chain — edges below confidence or score thresholds. Read-only.
  • truthsea_flag_weak_link — Flag a weak edge for the bounty program. Earn 100 TRUTH if the edge is later invalidated within 30 days. Requires wallet.

Chain Score Formula

chainScore = intrinsicScore * (floor + damping * weakestDepChainScore / 10000) / 10000

Where:

  • intrinsicScore = weighted average of 4 truth frameworks (30% correspondence, 25% coherence, 25% convergence, 20% pragmatism)
  • floor = 3000 (0.30 minimum preservation)
  • damping = 7000 (0.70 attenuation factor)
  • Contradiction penalty: -15% per contradicting edge (min 40% preservation)

Scoring System

Truth Frameworks (0-100 each)

  1. Correspondence — Maps to observable reality
  2. Coherence — Fits the web of known truths
  3. Convergence — Independent sources agree over time
  4. Pragmatism — Works in practice

Moral Vector (-100 to +100 each)

  1. Care / Harm
  2. Fairness / Cheating
  3. Loyalty / Betrayal
  4. Authority / Subversion
  5. Sanctity / Degradation
  6. Liberty / Oppression
  7. Epistemic Humility / Dogmatism
  8. Temporal Stewardship / Short-term Extraction

Commands

Truth Verification (V1)

CommandDescription
----------------------
/verify Submit a claim for multi-dimensional truth verification
/truth query Search verified truth quanta
/dispute Challenge a quantum with counter-evidence

Dependency Graph (V2)

CommandDescription
----------------------
/edge create Create a dependency edge between quanta
/edge dispute Challenge an edge — earn TRUTH if justified
/dag explore Navigate the dependency graph
/dag score Get the propagated chain score
/dag weak-links Find weak foundations in the dependency chain
/dag flag Flag a weak edge — earn 100 TRUTH bounty if validated

Bounties (CrowdedSea)

CommandDescription
----------------------
/bounty listList available truth bounties with ETH rewards
/bounty claim Claim a bounty for investigation

> Note: V2 DAG commands require TruthDAG contract address to be configured.

Token Incentives

ActionReward
----------------
Host a quantum100 TRUTH
Verify a quantum10 TRUTH
Create an edge (survives 7 days)20 TRUTH
Trigger score propagation2 TRUTH
Validated weak-link flag100 TRUTH
Win edge dispute60% of proposer's stake + 20 TRUTH

Contracts (Base Sepolia)

  • TruthToken: 0x18D825cE88089beFC99B0e293f39318D992FA07D
  • TruthRegistryV2: 0xbEE32455c12002b32bE654c8E70E876Fd557d653
  • BountyBridge: 0xA255A98F2D497c47a7068c4D1ad1C1968f88E0C5

Security (V2.5)

  • ReentrancyGuard on all ETH transfer functions
  • Per-era emission caps enforced (halving schedule)
  • Slither static analysis with 0 high/critical findings
  • API rate limiting, CORS whitelist, CSP headers

Links

版本历史

共 1 个版本

  • v2.5.1 当前
    2026-05-12 05:22 安全 安全

安全检测

腾讯云安全 (Keen)

安全,无风险
查看报告

腾讯云安全 (Sanbu)

安全,无风险
查看报告

🔗 相关推荐

ai-intelligence

Self-Improving + Proactive Agent

ivangdavila
自我反思+自我批评+自我学习+自组织记忆。智能体评估自身工作、发现错误并持续改进。
★ 1,353 📥 317,923
ai-intelligence

self-improving agent

pskoett
捕获经验教训、错误和纠正,以实现持续改进。使用时机:(1)命令或操作意外失败;(2)用户纠正……
★ 4,058 📥 797,615
security-compliance

Skill Vetter

spclaudehome
AI智能体技能安全预审工具。安装ClawdHub、GitHub等来源技能前,检查风险信号、权限范围及可疑模式。
★ 1,212 📥 266,341