name: 🤖 Report a Prompt Handling Failure
description: "File a report for a user prompt that the AI agent did not handle as expected"
labels: ["prompt-failure"]
body:
- type: markdown
attributes:
value: |
Thank you for taking the time to report an issue! This template will help us understand and reproduce the problem you're facing with the AI agent's interaction with the Blockscout MCP server.
> **⚠️ IMPORTANT: DO NOT INCLUDE SENSITIVE INFORMATION!**
>
> Please do not paste any private keys, API keys, or other confidential data. If you need to include addresses, transaction hashes, or other on-chain data from a private context, please replace them with placeholder values.
>
> **If you replace any part of your prompt or the agent's response, please clearly mark it.** For example, use `[REDACTED_ADDRESS]` or `[REPLACED_VALUE]`. This helps us understand that the provided data is not the literal original, and we will adjust our investigation accordingly.
- type: input
id: llm_client
attributes:
label: LLM Client
description: Which client are you using?
placeholder: "e.g., Claude Desktop, Cursor, MCP Inspector"
validations:
required: true
- type: input
id: llm_model
attributes:
label: LLM Model
description: Which model are you using?
placeholder: "e.g., Claude 3.5 Sonnet, GPT-4o, Llama 3"
validations:
required: true
- type: input
id: mcp_version
attributes:
label: MCP Server Version
description: Which version of the Blockscout MCP server are you using? (optional)
placeholder: "e.g., 0.3.1, latest Docker image"
- type: textarea
id: user_prompt
attributes:
label: User Prompt
description: Please paste the exact prompt you gave to the AI agent below
placeholder: |
<Your prompt here>
render: text
validations:
required: true
- type: textarea
id: agent_response
attributes:
label: Agent's Response / Behavior
description: |
Please describe the agent's full response or behavior (optional).
Include the final text answer, the sequence of tool calls made (if visible), and any error messages.
placeholder: |
<Paste the agent's response or describe its behavior here>
render: markdown
- type: textarea
id: expected_behavior
attributes:
label: Expected Behavior
description: What did you expect the agent to do? What would a correct response or sequence of tool calls look like? (optional)
- type: textarea
id: additional_context
attributes:
label: Additional Context
description: |
Add any other context about the problem here. You can include screenshots, logs, or further comments.
