README.md

---
pretty_name: "GLM-5.2 Agent traces"
task_categories:
- text-generation
tags:
- "agent-traces"
- "format:agent-traces"
- "pi"
- "distillation"
- "z-ai/glm-5.2"
- "teich"
configs:
- config_name: default
  data_files:
  - split: train
    path: "**/*.jsonl"
---

This dataset was generated using [teich](https://github.com/TeichAI/teich) by [TeichAI](https://huggingface.co/TeichAI) <img src="https://cdn-avatars.huggingface.co/v1/production/uploads/6837935ac3b7ffe0d2559ce9/-AxyvV4wfUY8uo87kNKkK.png" width="20" height="20" style="display: inline-block; vertical-align: middle; margin: 0 3px;">

# GLM-5.2 Agent traces

This directory contains raw agent trace files generated by teich.

JSONL files: 86

Model metadata: `z-ai/glm-5.2`

## Training-ready tools

Generated agent traces carry configured or recovered tool schemas so tools remain available for training even when a session did not call them.
Native Claude Code imports recover schemas for Claude Code and Claude Desktop built-ins, plus conservative name-derived MCP schemas, when the raw transcript only records tool names or calls.
A complete dataset-level `tools` schema snapshot is embedded in the collapsed section at the bottom of this README.
`load_traces` applies the dataset snapshot to each loaded example as a fallback `tools` field.

## Format

Each file is newline-delimited JSON representing a single captured agent session.
The trace schema is designed for upload-first preservation so you can keep the original session history and convert it later for training.
Teich normalizes split assistant fragments during trace copy and conversion so the semantic order is reasoning first, optional assistant text second, and tool calls last.
Native Claude Code conversion also preserves runtime context such as skills, MCP instructions, hook context, permission state, date changes, and session recaps as masked `system` messages when the raw transcript provides them.

Common top-level event groups:

- `session_meta`
- `turn_context`
- `event_msg`
- `response_item`
- `session`
- `message`
- `session_info`
- `model_change`
- `thinking_level_change`
- `external_session_meta`
- `external_message`
- `external_stderr`

## Example

```json
{"type": "session", "version": 3, "id": "019ee172-2f9f-7341-8125-4e5d535d9e56", "timestamp": "2026-06-19T19:53:37.440Z", "cwd": "/workspace"}
```

## Training

Use this dataset as `AletheiaResearch/GLM-5.2-Agent` with Teich's data preparation and training utilities.
If you do not want Teich to handle chat-template formatting or masking, run `teich convert` to write standalone OpenAI-style JSONL rows with `prompt`, `messages`, `tools`, and `metadata`.
Training setup details evolve over time, so the maintained guide lives in the [Teich training docs](https://github.com/TeichAI/teich/blob/main/docs/training.md).
For loading, mixing, converting, and validating Teich datasets, see [Preparing Data](https://github.com/TeichAI/teich/blob/main/docs/prepare-data.md).

## Tool schema snapshot

<details>
<summary>Training-ready tool schema snapshot</summary>

```json
[
  {
    "type": "function",
    "function": {
      "name": "bash",
      "description": "Run shell commands in the workspace.",
      "parameters": {
        "type": "object",
        "properties": {
          "cmd": {
            "type": "string"
          },
          "command": {
            "type": "string"
          },
          "cwd": {
            "type": "string"
          },
          "description": {
            "type": "string"
          },
          "timeout": {
            "type": "integer"
          }
        },
        "additionalProperties": true
      }
    }
  },
  {
    "type": "function",
    "function": {
      "name": "edit",
      "description": "Edit file contents in the workspace.",
      "parameters": {
        "type": "object",
        "properties": {
          "edits": {
            "type": "array"
          },
          "file_path": {
            "type": "string"
          },
          "path": {
            "type": "string"
          }
        },
        "additionalProperties": true,
        "required": [
          "edits"
        ]
      }
    }
  },
  {
    "type": "function",
    "function": {
      "name": "read",
      "description": "Read file contents from the workspace.",
      "parameters": {
        "type": "object",
        "properties": {
          "file_path": {
            "type": "string"
          },
          "limit": {
            "type": "integer"
          },
          "offset": {
            "type": "integer"
          },
          "path": {
            "type": "string"
          }
        },
        "additionalProperties": true
      }
    }
  },
  {
    "type": "function",
    "function": {
      "name": "read_file",
      "description": "Read file contents from the workspace.",
      "parameters": {
        "type": "object",
        "properties": {
          "path": {
            "type": "string"
          }
        },
        "additionalProperties": true,
        "required": [
          "path"
        ]
      }
    }
  },
  {
    "type": "function",
    "function": {
      "name": "write",
      "description": "Write file contents in the workspace.",
      "parameters": {
        "type": "object",
        "properties": {
          "content": {
            "type": "string"
          },
          "file_path": {
            "type": "string"
          },
          "path": {
            "type": "string"
          }
        },
        "additionalProperties": true,
        "required": [
          "content"
        ]
      }
    }
  },
  {
    "type": "function",
    "function": {
      "name": "write_file",
      "description": "Write file contents in the workspace.",
      "parameters": {
        "type": "object",
        "properties": {
          "content": {
            "type": "string"
          },
          "path": {
            "type": "string"
          }
        },
        "additionalProperties": true,
        "required": [
          "content",
          "path"
        ]
      }
    }
  }
]
```

</details>