multi-sig-account-permissions
Manage TRON multi-sig permissions — configure keys, thresholds, and co-signed proposals.
适合你,如果需要管理 TRON 区块链上的多签账户和权限
npx oh-my-skill add bofai/skills/multi-sig-account-permissionscurl -fsSL https://oh-my-skill.com/install.sh | bash -s -- bofai/skills/multi-sig-account-permissionsnpx oh-my-skill verify bofai/skills/multi-sig-account-permissions怎么用
商店整理自技能原文 · 版本 39919d1 · 表述以原文为准装上后,你可以管理TRON账户的权限:查看和修改权限设置,创建、批准和执行多签交易提案。
当你提出管理TRON账户权限、设置多签或处理多签提案时触发。
技能原文 SKILL.md
Multi-Sig & Account Permissions
Manage TRON's native three-tier permission model (Owner, Active, Witness) with built-in multi-sig support at the protocol level — no external contracts needed. Configure keys, thresholds, and operation scopes, then coordinate multi-signature transactions through a propose → approve → execute flow.
Quick Start
Wallet required: Runagent-wallet listfirst. If no wallets exist, invokebankofai-guide(Section C — Wallet Guard) before proceeding.
cd multisig-permissions && npm install export TRON_PRIVATE_KEY="<your-private-key>" export TRON_NETWORK="mainnet"
Before using any signer-specific command, derive the address from the configured key and confirm it matches the intended owner / active role on-chain. Variable names like TRON_PRIVATE_KEY and TRON_HUMAN_PRIVATE_KEY are only labels; the actual signer role is determined by the derived address in the current account permission set.
[!CAUTION] Misconfiguring owner permissions can permanently lock an account with no recovery. Always use --dry-run first and verify you control enough keys to meet the new threshold.
Available Scripts
| Script | Purpose | Reads/Writes | |--------|---------|-------------| | status.js | View current permission configuration | Read-only | | update.js | Modify account permissions (add/remove keys, thresholds, scope) | Write | | propose.js | Create a multi-sig transaction proposal | Write | | approve.js | Add your signature to a pending proposal | Write | | execute.js | Broadcast a fully-signed transaction | Write | | pending.js | List pending multi-sig proposals | Read-only | | review.js | Human CLI: list, inspect, co-sign, and execute proposals in one tool | Write |
TRON Permission Model
Every TRON account has three permission tiers:
| Permission | Type | Purpose | Default | |-----------|------|---------|---------| | Owner | 0 | Full control. Only permission that can modify other permissions. | Account key, threshold 1 | | Active | 2 | Day-to-day operations. Can be scoped to specific transaction types. | Account key, threshold 1, all ops | | Witness | 1 | Block production (Super Representatives only). | Not set |
Each permission has keys (address + weight pairs) and a threshold (minimum total weight to authorize). Example: 3 keys with weight 1 each and threshold 2 = 2-of-3 multi-sig.
Usage Patterns
Pattern 1: Inspect Account Security
# Check your own account node scripts/status.js # Check any account node scripts/status.js TXk8rQSAvPvBBNtqSoY3UkFdpMTMbqRMKU
Pattern 2: Set Up Multi-Sig (2-of-3)
# Step 1: Preview changes (always dry-run first!) node scripts/update.js add-key TKey2Address... --permission owner --weight 1 --dry-run node scripts/update.js add-key TKey3Address... --permission owner --weight 1 --dry-run node scripts/update.js set-threshold 2 --permission owner --dry-run # Step 2: Or use a template for all at once node scripts/update.js from-template basic-2of3 \ --key1 TKey1... --key2 TKey2... --key3 TKey3... --dry-run # Step 3: Execute when satisfied node scripts/update.js from-template basic-2of3 \ --key1 TKey1... --key2 TKey2... --key3 TKey3...
Pattern 3: Restrict Agent to DeFi Only
# Preview the restricted owner + active configuration node scripts/update.js from-template agent-restricted \ --key1 THumanKey... --key2 TBackupKey... --key3 TAgentKey... --dry-run # Apply it from a single-owner account node scripts/update.js from-template agent-restricted \ --key1 THumanKey... --key2 TBackupKey... --key3 TAgentKey... # If the current owner is already multi-sig, update.js creates a pending owner # proposal instead of broadcasting directly. Finish it through the normal flow. node scripts/approve.js prop_xxxxx_xxxx node scripts/execute.js prop_xxxxx_xxxx
Pattern 4: Multi-Sig Transaction Flow
# Signer 1: Propose a transfer node scripts/propose.js transfer TRecipient... 10000 --memo "Q1 budget" # Signer 2: Review and approve node scripts/approve.js prop_1709312400_a3f2 # Any signer: Execute when threshold is met node scripts/execute.js prop_1709312400_a3f2 # Check all pending proposals node scripts/pending.js
Pattern 5: Hybrid Signature (Human + Agent)
The review.js script is a single CLI tool for humans to list, inspect, co-sign, and execute agent-created proposals. It uses TRON_HUMAN_PRIVATE_KEY (not TRON_PRIVATE_KEY) to avoid mixing up human and agent keys.
# Agent (Key B) proposes a transfer from its own account (uses TRON_PRIVATE_KEY): node scripts/propose.js transfer TRecipient... 500 --memo "Q1 vendor payment" # → Outputs proposal ID: prop_1710345600_b7e4 # Agent (Key B) proposes an active-scoped contract call for a controlled # multi-sig account. --account identifies the controlled account. node scripts/propose.js contract-call TXLAQ63Xg1NAzckPwKHvzw7CSEmLMEqcdj \ "approve(address,uint256)" '["TSpender...","1000000"]' \ --permission active \ --account TControlledMultisig... # Human (Key A) reviews all pending proposals (no key needed): node scripts/review.js # Human inspects a specific proposal (read-only, no key needed): node scripts/review.js prop_1710345600_b7e4 # Human co-signs after verifying details (uses TRON_HUMAN_PRIVATE_KEY): export TRON_HUMAN_PRIVATE_KEY="<human-key>" node scripts/review.js prop_1710345600_b7e4 --sign # Human co-signs AND broadcasts in one step: node scripts/review.js prop_1710345600_b7e4 --sign --execute
Pattern 6: Scope Active Permission Operations
# Restrict active permission to only TRX transfers and smart contract calls node scripts/update.js scope-active --id 2 \ --operations TransferContract,TriggerSmartContract --dry-run
Permission Templates
Templates use --keyN flags to assign addresses to named roles. Keys are numbered positionally across owner and active roles (deduplicated, owner first). You can also use --role_name directly (e.g. --agent_key TAddr...).
| Template | Config | Key mapping | Description | |----------|--------|-------------|-------------| | basic-2of3 | 2-of-3 owner, 1-of-1 active (all ops) | --key1=KEY_1, --key2=KEY_2, --key3=KEY_3 | Standard multi-sig | | agent-restricted | 2-of-2 owner, 1-of-1 active (TriggerSmartContract only) | --key1=HUMAN_KEY, --key2=BACKUP_KEY, --key3=AGENT_KEY | Agent can only call contracts | | team-tiered | 3-of-5 owner, 2-of-3 active (transfers + contracts) | --key1..--key5=KEY_1..KEY_5 | Team with tiered access | | weighted-authority | Threshold 3: primary (wt 2) + secondaries (wt 1) | --key1=PRIMARY_KEY, --key2=SECONDARY_1, --key3=SECONDARY_2 | Weighted key authority |
Transaction Types (Operations)
Used with scope-active --operations:
| Operation | Description | |-----------|-------------| | TransferContract | Transfer TRX | | TransferAssetContract | Transfer TRC10 tokens | | TriggerSmartContract | Call any smart contract (DeFi, tokens, etc.) | | FreezeBalanceV2Contract | Stake TRX for energy/bandwidth | | UnfreezeBalanceV2Contract | Unstake TRX | | DelegateResourceContract | Delegate energy/bandwidth | | VoteWitnessContract | Vote for Super Representatives | | AccountPermissionUpdateContract | Modify permissions (very dangerous in Active) |
See resources/permission_config.json for the full list.
Security Rules
[!WARNING] Lockout prevention: The update.js script refuses to execute if the new threshold exceeds the total key weight. However, you MUST verify you actually control the keys — the script cannot check this.
- Always dry-run first: Every write operation supports
--dry-run. - Owner changes are irreversible: Only the owner key set can modify permissions. If you lock yourself out, the account is permanently lost.
- Verify key control: Before changing owner permissions, confirm you have access to enough keys to meet the new threshold.
- Scope active permissions: Don't give agents all-operations access. Use
scope-activeto limit toTriggerSmartContract. - Proposals expire: Default expiry is 24 hours. Expired proposals cannot be executed.
- Check signer roles explicitly: Before
approve.js,review.js --sign, or any active-scopedpropose.js --accountcall, confirm the configured private key derives to an address that actually appears in that permission block.
Multi-Sig Proposal Storage
Proposals are stored locally at ~/.clawdbot/multisig/pending/ as JSON files. After execution, they are archived to ~/.clawdbot/multisig/executed/.
For multi-party setups, share the proposal ID with co-signers who have access to the same file system, or copy the proposal JSON file to them.
Common Issues
"LOCKOUT DANGER: threshold exceeds total key weight"
The script prevented a dangerous permission change. Reduce the threshold or add more keys.
"Account not found or not activated"
The address has no on-chain account. Send at least 0.1 TRX to activate it.
"Proposal has expired"
The proposal's 24-hour window has passed. Create a new proposal with propose.js.
"Threshold not met"
The proposal needs more signatures. Check pending.js to see remaining weight needed.
Environment Variables
| Variable | Required | Description | |----------|----------|-------------| | TRON_PRIVATE_KEY | Yes (write scripts) | Private key used by propose.js, approve.js, execute.js, and update.js. The derived address must match the intended owner or active signer role. | | TRON_HUMAN_PRIVATE_KEY | Yes (review.js --sign) | Private key used by review.js --sign. This should be set to the human reviewer's key and validated by derived address, not by variable name alone. | | TRON_NETWORK | No (default: mainnet) | mainnet, nile, or shasta | | TRONGRID_API_KEY | No | TronGrid API key for higher rate limits |
Version 1.0.0 — Created by M2M Agent Registry for Bank of AI