Skip to content

Commit ac4962e

Browse files
committed
revise agent documentation: streamline sections on agent editing, registration, and root agents for enhanced user understanding and experience
1 parent 6ce89fd commit ac4962e

File tree

3 files changed

+28
-278
lines changed

3 files changed

+28
-278
lines changed

src/content/docs/explanations/agent-editing.mdx

Lines changed: 9 additions & 133 deletions
Original file line numberDiff line numberDiff line change
@@ -12,118 +12,23 @@ import {
1212
TabItem,
1313
} from "@astrojs/starlight/components";
1414

15-
This explanation covers the agent management system in Torus - how agent ownership works, what aspects of agents can be modified, and the underlying mechanics of agent updates.
15+
## Agent Editing System
1616

17-
## Agent Ownership and Permissions
17+
Agent editing in Torus allows agents to update their metadata, endpoints, and social connections after initial registration. Only the original registering wallet maintains editing rights, ensuring agent integrity.
1818

19-
Agent management in Torus follows a strict ownership model where only the wallet address that originally registered an agent maintains modification rights. This ensures agent integrity and prevents unauthorized changes to agent profiles.
19+
Agents can modify display information (titles, descriptions), service endpoints, visual assets, and social media links. Core identifiers and cryptographic associations remain immutable once set during registration.
2020

21-
### Modifiable Agent Properties
21+
### Update Economics
2222

23-
Agents contain both **immutable** and **mutable** properties:
23+
Updates require only standard transaction fees - no token burning like initial registration. This minimal cost structure encourages agents to maintain current, accurate information while preventing spam through basic network fees.
2424

25-
**Immutable Properties** (Set at registration)
26-
- Agent identifier/name
27-
- Initial registration block
28-
- Cryptographic key associations
25+
Regular updates improve discoverability and signal ongoing agent activity, building trust within the Torus ecosystem.
2926

30-
**Mutable Properties** (Can be updated)
31-
- Display metadata (title, descriptions)
32-
- Service endpoints and URLs
33-
- Visual assets and branding
34-
- Social media connections
27+
### Storage Architecture
3528

36-
## Update Transaction Mechanics
29+
Agent information uses a hybrid model: core identifiers and ownership on-chain, with extended metadata and rich content stored off-chain via IPFS. This balances blockchain immutability with the flexibility needed for comprehensive agent profiles.
3730

38-
Agent updates are implemented as blockchain transactions that modify the agent's on-chain metadata. The update process involves:
39-
40-
1. **Ownership Verification**: The network verifies the transaction signer owns the agent
41-
2. **Metadata Validation**: New metadata is validated against schema requirements
42-
3. **State Transition**: The agent's on-chain state is updated atomically
43-
4. **Event Emission**: Update events are broadcast for indexing and discovery
44-
45-
## Agent Metadata Architecture
46-
47-
Agent information is stored using a hybrid on-chain/off-chain model:
48-
49-
### On-Chain Storage
50-
- Core identifiers and ownership
51-
- Service endpoint URLs
52-
- Cryptographic keys and signatures
53-
- Update timestamps and version history
54-
55-
### Off-Chain Storage (IPFS)
56-
- Extended metadata and descriptions
57-
- Visual assets and media files
58-
- Social media links and external references
59-
- Rich content and documentation
60-
61-
This architecture balances blockchain immutability with the flexibility needed for rich agent profiles.
62-
63-
### Metadata Schema Structure
64-
65-
Agent metadata follows a standardized JSON schema:
66-
67-
<Code
68-
code={`{
69-
"name": "agent_identifier",
70-
"title": "Agent Display Name",
71-
"short_description": "Brief description (max 200 characters)",
72-
"description": "Detailed description of your agent's capabilities and purpose (max 5000 characters)",
73-
"website_url": "https://example.com",
74-
"api_endpoint_url": "https://api.example.com",
75-
"agent_icon": "uploaded_image_file",
76-
"social_links": {
77-
"discord": "https://discord.gg/example",
78-
"twitter": "https://x.com/example",
79-
"github": "https://github.com/example",
80-
"telegram": "https://t.me/example"
81-
}
82-
}`}
83-
lang="json"
84-
title="agent-fields.json"
85-
/>
86-
87-
**Schema Constraints:**
88-
89-
- **Character Limits**: Enforce reasonable bounds on text fields
90-
- **URL Validation**: Ensure endpoint accessibility and format compliance
91-
- **Image Requirements**: Optimize for consistent visual presentation
92-
- **Social Link Verification**: Validate platform-specific URL formats
93-
94-
## Economic Model for Updates
95-
96-
Agent updates follow a different economic model than initial registration:
97-
98-
### Fee Structure
99-
- **Transaction Costs Only**: Updates only require standard network transaction fees
100-
- **No Burn Mechanism**: Unlike registration, updates don't permanently remove tokens from circulation
101-
- **Dynamic Fee Pricing**: Costs fluctuate based on network congestion and demand
102-
103-
### Economic Rationale
104-
The minimal update cost structure encourages agents to maintain accurate, current information while preventing spam through basic transaction fees.
105-
106-
## Agent Lifecycle Considerations
107-
108-
### Maintenance Strategy
109-
Successful agents typically maintain regular update cycles to:
110-
- **Reflect Capability Evolution**: Update descriptions as services expand
111-
- **Maintain Discoverability**: Keep social links and contact information current
112-
- **Signal Activity**: Regular updates demonstrate ongoing agent operation
113-
114-
### Version Control and History
115-
The blockchain maintains a complete history of agent updates, enabling:
116-
- **Audit Trails**: Track changes over time for accountability
117-
- **Rollback Capabilities**: Reference previous states if needed
118-
- **Reputation Building**: Demonstrate consistent, professional maintenance
119-
120-
### Torus Discovery Impact
121-
Agent updates affect Torus-wide discovery and matching:
122-
- **Search Indexing**: Updated metadata improves discoverability
123-
- **Capability Matching**: Current descriptions enable better agent-to-agent connections
124-
- **Trust Signals**: Well-maintained profiles build confidence in agent reliability
125-
126-
## Related Concepts
31+
### Related Concepts
12732

12833
- **[Agent Registration](https://docs.torus.network/explanations/agent-registration/)** - Initial agent creation and ownership establishment
12934
- **[Demand Signaling](https://docs.torus.network/explanations/demand-signaling/)** - How agents coordinate through signals
@@ -132,32 +37,3 @@ Agent updates affect Torus-wide discovery and matching:
13237
<Aside type="tip" title="Ready to Update Your Agent?">
13338
Follow our [step-by-step guide to edit your agent](https://docs.torus.network/how-to-guides/edit-your-agent/) for practical instructions.
13439
</Aside>
135-
136-
---
137-
138-
## Deprecated CLI Method
139-
140-
<Aside type="caution" title="Deprecated">
141-
The CLI method for updating agents is deprecated and should only be used for testing or development purposes. We strongly recommend using the web application instead.
142-
</Aside>
143-
144-
### Command Line Interface (Legacy)
145-
146-
```bash
147-
torus agent update <name> <key> <url> <CID>
148-
```
149-
150-
**Parameters:**
151-
- `name`: Your agent's current identifier
152-
- `key`: Your agent's cryptographic key
153-
- `url`: Updated service endpoint URL
154-
- `CID`: IPFS Content Identifier for your updated metadata
155-
156-
**Requirements for CLI updates:**
157-
- Updated agent JSON metadata following the schema above
158-
- IPFS access to upload and pin metadata files
159-
- Direct access to agent's cryptographic keys
160-
161-
<Aside type="caution">
162-
CLI updates require manual IPFS management and have higher risk of errors. Use at your own risk.
163-
</Aside>

src/content/docs/explanations/agent-registration.mdx

Lines changed: 11 additions & 93 deletions
Original file line numberDiff line numberDiff line change
@@ -19,59 +19,24 @@ import {
1919
TabItem,
2020
} from "@astrojs/starlight/components";
2121

22-
This reference explains the technical aspects of agent registration, including costs, metadata requirements, and the underlying mechanisms that power agent discovery and interaction in Torus.
22+
## Agent Registration System
2323

24-
### What is a Agent?
24+
Agent registration in Torus creates discoverable entities that can receive, create, and delegate permissions within the network. Registration is immediate with no approval required, but involves burning tokens to prevent spam.
2525

26-
**Agent Features:**
26+
### Registration Economics
2727

28-
- Can receive, create and delegate permissions & constraints
29-
- Simple registration process with no approval required
30-
- Immediate activation after registration
28+
Agent registration requires permanently burning tokens:
3129

30+
- **Agent Registration Fee**: Fixed 15 TORUS fee (burned permanently)
31+
- **Capability Creation Fee**: Fixed 3.29 TORUS fee (burned permanently)
32+
- **Capability Deposit**: Variable deposit based on agent name size (refundable)
33+
- **Transaction Fee**: Variable network fee (burned permanently)
3234

33-
### Registration Costs
35+
The burn mechanism maintains network quality by creating economic barriers to spam while allowing the capability deposit to be recovered when removed.
3436

35-
<Aside type="note">
36-
Registration requires burning tokens permanently. The amount varies based on
37-
network conditions and will be displayed before you confirm the transaction.
38-
</Aside>
39-
40-
#### Fee Breakdown
41-
42-
Registration fees are divided into several components:
43-
44-
- **Transaction Fee**: Variable network fee. Burned permanently
45-
- **Agent Registration Fee**: Fixed 15 TORUS fee. Burned permanently
46-
- **Capability Creation Fee**: Fixed 3.29 TORUS fee. Burned permanently
47-
- **Capability Deposit**: Variable deposit based on the size of the agent name. This value is refunded when the capability is removed
48-
49-
#### Understanding the Burn Mechanism
50-
51-
- **Permanent Loss**: Most fees are burned and cannot be recovered (except capability deposit)
52-
- **Purpose**: Prevents spam and maintains network quality
53-
- **Refundable Component**: Only the capability deposit can be recovered by removing the capability
54-
55-
### Agent Discovery and Interaction
56-
57-
Once registered, agents become discoverable through:
58-
59-
- **Agent Registry**: Public listing of all registered agents
60-
- **Metadata Resolution**: Automatic fetching and display of agent information
61-
- **Permission System**: Ability to receive and delegate permissions
62-
- **Capability Registration**: Creating named capabilities for other agents to invoke
63-
64-
### Technical Implementation
65-
66-
#### On-Chain Storage
67-
- Agent registration creates an on-chain record linking your address to metadata
68-
- Metadata is stored off-chain (IPFS) but referenced on-chain via content hash
69-
- Registration transaction burns tokens permanently to prevent spam
37+
### Discovery and Capabilities
7038

71-
#### Torus Integration
72-
- Registered agents can participate in governance proposals
73-
- Agents can create and manage capabilities within their namespace
74-
- Other agents can discover and interact with registered agents through the protocol
39+
Registered agents become discoverable through the public agent registry and can be viewed on the [Allocator page](https://allocator.torus.network/). They can participate in the permission system, create named capabilities for other agents to invoke, and engage in governance proposals. Agent metadata is stored off-chain via IPFS but referenced on-chain through content hashes.
7540

7641
### Related Concepts
7742

@@ -82,50 +47,3 @@ Once registered, agents become discoverable through:
8247
<Aside type="tip" title="Ready to Register?">
8348
Follow our [step-by-step registration guide](https://docs.torus.network/how-to-guides/register-an-agent/) to register your agent.
8449
</Aside>
85-
86-
---
87-
88-
## Deprecated: CLI Registration Method
89-
90-
<Aside type="caution" title="Deprecated">
91-
The CLI registration method is deprecated and no longer recommended. Use the web application instead.
92-
</Aside>
93-
94-
### Agent Metadata Schema (CLI Only)
95-
96-
<Aside type="note">
97-
This metadata schema was used for CLI-based registration which is now deprecated. The web application handles all metadata automatically.
98-
</Aside>
99-
100-
<Code
101-
code={`{
102-
"title": "Agent Display Name",
103-
"short_description": "Brief description (max 200 characters)",
104-
"description": "Detailed description of your agent's capabilities and purpose",
105-
"website": "https://example.com",
106-
"images": {
107-
"icon": "ipfs://bafybeif6b2le2hd4cwceqsxdvtugqrjshsuwuy64r6c7pvqwltbezrhpry",
108-
"banner": "ipfs://bafybeif6b2le2hd4cwceqsxdvtugqrjshsuwuy64r6c7pvqwltbezrhpry"
109-
},
110-
"socials": {
111-
"discord": "https://discord.com/invite/example",
112-
"github": "https://github.com/username",
113-
"telegram": "https://t.me/example",
114-
"twitter": "https://twitter.com/username"
115-
}
116-
}`}
117-
lang="json"
118-
title="agent-metadata.json"
119-
/>
120-
121-
#### Metadata Field Requirements (CLI Only)
122-
123-
| Field | Type | Required | Description |
124-
| ------------------- | -------- | -------- | ------------------------------------ |
125-
| `title` | string | Yes | Display name for your agent |
126-
| `short_description` | string | Yes | Brief summary (max 200 characters) |
127-
| `description` | string | Yes | Detailed explanation of capabilities |
128-
| `website` | string | No | Agent's primary website URL |
129-
| `images.icon` | string | No | IPFS hash for agent icon |
130-
| `images.banner` | string | No | IPFS hash for agent banner |
131-
| `socials.*` | string | No | Social media and contact links |

src/content/docs/explanations/root-agents.mdx

Lines changed: 8 additions & 52 deletions
Original file line numberDiff line numberDiff line change
@@ -19,63 +19,19 @@ import {
1919
TabItem,
2020
} from "@astrojs/starlight/components";
2121

22-
This reference explains the Root Agent approval process, DAO evaluation criteria, and the strategic considerations for agents seeking direct emission access.
22+
## Root Agent System
2323

24-
### What Makes a Root Agent
24+
Root Agents are DAO-approved agents that receive direct emissions from the stake root and can set network goals around which swarms of agents organize. Unlike standard agents, Root Agents access funding without intermediaries and guide network development in areas of unique expertise.
2525

26-
Root Agents are distinguished from Registered Agents by their ability to:
26+
### DAO Approval Process
2727

28-
- **Receive Direct Emissions**: Access funding directly from the stake root without intermediaries
29-
- **Set Torus Goals**: Establish objectives around which swarms of agents can organize
30-
- **Lead Coordination**: Guide network development in areas of unique expertise
28+
Becoming a Root Agent requires DAO member approval through [whitelist applications](https://dao.torus.network/whitelist-applications). Applications need majority approval and are evaluated based on technical capability, unique value proposition, and community standing.
3129

32-
Root Agent status is reserved for agents with clear, unique value propositions that don't overlap with existing Root Agents.
30+
Successful applicants demonstrate credible execution, active community participation, and non-overlapping scope that complements existing Root Agents.
3331

34-
## What to Expect from a Root Agent
35-
36-
### Technical Excellence
37-
- **Credible Execution**: Signals (code, clarity, discussion) that inspire confidence in your technical ability
38-
- **Technical Innovation**: Novel approaches or solutions that advance the network
39-
40-
### Community Standing
41-
- **Active Participation**: Regular engagement in Discord, governance, and development discussions
42-
- **Transparent Communication**: Clear, honest communication about goals and progress
43-
44-
### Unique Value Proposition
45-
- **Non-Overlapping Scope**: Services that complement rather than duplicate existing Root Agents
46-
- **Torus Benefit**: Clear explanation of how the agent improves Torus for all participants
47-
48-
## DAO Voting Process
49-
50-
### How Voting Works
51-
52-
- **DAO Members Vote**: Existing DAO members review and [vote on whitelist applications](https://dao.torus.network/whitelist-applications)
53-
- **Majority Required**: Applications need majority approval to pass
54-
55-
### Factors DAO Members Consider
56-
57-
- **Technical Capability**: Does the agent provide genuine utility?
58-
- **Unique Value**: What unique services or capabilities does this agent offer?
59-
60-
## Application Tips
61-
62-
### Increasing Your Approval Chances
63-
64-
**Before Applying:**
65-
66-
- Demonstrate technical competency
67-
- Be active in the community Discord
68-
69-
**In Your Application:**
70-
71-
- Provide detailed, specific information about your agent's capabilities
72-
- Explain clearly what unique value you bring to the network
73-
74-
### Financial Considerations
75-
76-
#### Application Costs
77-
- **Initial Fee**: 100 TORUS is required when submitting your application
78-
- **Refund Policy**: If approved, the fee is refunded. If rejected, the fee is burned permanently.
32+
<Aside type="note">
33+
Applications require a 100 TORUS fee. If approved, the fee is refunded. If rejected, the fee is burned permanently.
34+
</Aside>
7935

8036
### Related Concepts
8137

0 commit comments

Comments
 (0)