Query address info Skill from Binance

[markusha77]

New Skill Contribution

Skill Info

• Skill ID: query-address-info
• Category: Blockchain
• Description: Query on-chain wallet addresses for token balances and positions, including token metadata, price, 24h change, and holding quantity.

Checklist

• SKILL.md has valid YAML frontmatter with all required fields
• Skill ID matches the directory name
• Category is one of: productivity, development, communication, writing, research, other
• Instructions are clear and self-contained
• Instructions are under 4000 tokens
• No external URLs referenced in instructions
• Tested the skill with an AI agent and it produces good results
• Added usage examples in the frontmatter

Testing

Describe how you tested this skill:

Notes

Any additional context or notes for reviewers.

[greptile-apps]

Greptile Summary

This PR adds a new query-address-info skill that documents the Binance Web3 API for querying on-chain wallet token balances and positions. While the API documentation itself is well-structured, the skill has two blocking issues that will fail automated CI checks before it can be merged, plus several functional gaps.

Key issues found:

• Invalid category value: Blockchain is not an allowed category. Valid values are productivity, development, communication, writing, research, other. This will fail the automated frontmatter validation.
• External URLs in instructions: The skill body embeds https://web3.binance.com/... and a full curl example, violating the "no external URLs" guideline. The PR checklist incorrectly marks this item as complete.
• No agent behavioral instructions: The file is API reference documentation only — there are no natural language instructions telling the agent how to respond, format output, handle errors, or decide when to paginate. Other skills in the repo (e.g., ab-test-setup) demonstrate the expected instruction-first format.
• Solana address format undocumented: Solana uses base58 public keys, not 0x hex addresses, but no guidance or example is provided for Solana wallet queries.
• Pagination termination condition missing: No total or hasMore field is documented in the response, and there is no limit parameter documented, leaving agents with no way to know when they have fetched all pages.
• Skill not tested: The PR checklist item for testing with an AI agent is unchecked.

Confidence Score: 1/5

• Not safe to merge — two P0 issues will fail automated CI checks, and the skill lacks the behavioral instructions needed to function as intended.
• The category: Blockchain value is not in the allowed list and will be rejected by the automated validator. The external URL in the instructions body also violates a checked guideline and will fail CI. Beyond CI failures, the skill is missing the core natural language agent instructions that define how skills in this repo work. These are blocking issues that require changes before the PR can be merged.
• skills/query-address-info/SKILL.md requires significant revision — invalid category, external URLs, and missing behavioral instructions all need to be addressed.

Important Files Changed

Filename	Overview
skills/query-address-info/SKILL.md	New skill adding Binance Web3 wallet address querying. Has two P0 issues (invalid category value, external URLs in instructions body) that will fail automated CI checks, a missing agent behavioral instructions section, an undocumented Solana address format incompatibility, and incomplete pagination documentation.

Sequence Diagram

sequenceDiagram
    participant User
    participant Agent
    participant BinanceWeb3 as Binance Web3 API

    User->>Agent: "Show token holdings for wallet 0x... on BSC"
    Agent->>Agent: Extract address + chainId (56)
    Agent->>BinanceWeb3: GET /active-position-list?address=0x...&chainId=56&offset=0
    Note over Agent,BinanceWeb3: Headers: clienttype: web, clientversion: 1.2.0
    BinanceWeb3-->>Agent: { code: "000000", data: { list: [...], offset: 0 } }
    Agent->>Agent: Check if more pages exist (undocumented)
    opt More pages available
        Agent->>BinanceWeb3: GET /active-position-list?address=0x...&chainId=56&offset=N
        BinanceWeb3-->>Agent: Next page of results
    end
    Agent->>Agent: Format results (price, percentChange24h, remainQty)
    Agent-->>User: Formatted portfolio summary

Loading

_{Last reviewed commit: "add Query Address In..."}

[greptile-apps]

Invalid category value

The category field is set to Blockchain, which is not one of the allowed values. Per CONTRIBUTING.md, valid categories are strictly: productivity, development, communication, writing, research, other. The automated CI checks validate this field and will reject the PR with this value. other is the correct fallback for blockchain/crypto skills.

Suggested change

	category: Blockchain
	category: other

[greptile-apps]

External URL referenced in skill instructions

The skill embeds a full external URL (https://web3.binance.com/...) directly in the instructions body. According to CONTRIBUTING.md, guideline #3 states: "Don't reference external URLs — Skills should be self-contained." This will also fail the automated PR checks.

The PR checklist item "No external URLs referenced in instructions" is incorrectly checked as complete. The URL, example curl command (line 51–55), and any other https:// references need to be replaced with instructional prose describing the endpoint path structure, so the agent can construct the request using knowledge already in the skill without reaching out to an external resource definition.

[greptile-apps]

Skill contains only API docs — no agent behavioral instructions

Looking at the rest of the skill repository (e.g., skills/ab-test-setup/SKILL.md), skills are written as natural language instructions that tell the AI agent how to behave. This file is purely API reference documentation (endpoint, parameters, response shape) with no instructions on what the agent should actually do.

There is no guidance such as:

• When should this skill activate?
• How should results be presented to the user (table, prose, list)?
• How should the agent handle pagination (when to fetch more pages, when to stop)?
• What should the agent say when the address has no holdings or when the API returns an error?
• How should the agent validate the wallet address format before calling the API?

Without behavioral instructions, an AI agent using this skill won't know how to respond meaningfully to a user request. Please add a natural language instructions section above the API documentation.

_{Note: If this suggestion doesn't match your team's coding style, reply to this and let me know. I'll remember it for next time!}

[greptile-apps]

Solana address format is undocumented and incompatible with the example request

The skill supports Solana (CT_501) but provides no guidance on its address format. Solana wallet addresses are base58-encoded public keys (e.g., 5YNmS1R9nNSCDzb5a7mMJ1dwK9uHeAAF4CmPEwKgVWr8), which are entirely different from the EVM 0x-prefixed hex addresses shown in all the examples. An agent following the examples would incorrectly format a Solana address query.

Please add a note clarifying the expected address format for Solana versus EVM chains, and provide a Solana-specific example request so agents can handle it correctly.

[greptile-apps]

Missing limit parameter and pagination termination condition

The offset parameter is documented but there is no limit parameter listed, and the response schema contains no total or hasMore field to signal when to stop paginating. An agent asked to retrieve all holdings would have no way to determine when the last page has been reached.

Consider:

1. Documenting any limit parameter if it exists.
2. Adding a field to the response table (e.g., total, hasNextPage) that indicates the total count or whether more pages are available.
3. Adding a note in the instructions on the recommended pagination strategy (e.g., "stop when list length is less than the page size").