Teamwork Graph CLI

https://twgcli.com
Admin
A CLI tool that unifies Atlassian products into a single knowledge graph with 380+ commands designed...
Tokens:1,782
Snippets:35
Trust Score:2.4
Update:4 months ago
Show doc for...
Tokens:
Raw
### Install TWG CLI

Source: https://twgcli.com/api

Installs the Teamwork Graph CLI using a curl command.

```bash
curl -fsSL https://twg.dev/install | sh
```

--------------------------------

### Example of concise Agent-friendly search

Source: https://twgcli.com/api

Demonstrates a search command with options for JSON output, automatic summary, and compact agent fields, piped to jq for specific field extraction.

```bash
$ twg search "incident review" --limit 20 \
    --output json \
    --output-summary auto \
    --agent-fields @compact \
    | jq '.items[] | {title, url, app}'

```

--------------------------------

### Example: Generating a sprint report

Source: https://twgcli.com

This example shows how an AI Agent can generate a sprint report using TWG CLI commands.

```bash
// 1. Resolve "mobile sprint 14"
$ twg resolve --query "mobile sprint 14"
→ atlassian:cloud/jira/sprint/14

// 2. Fetch issue and PR progress
$ twg jira sprint --id 14 --output json
$ twg bitbucket pull-requests --since 1w

// 3. Integrate data
$ twg context jira workitem MOBL-201 -o json | twg visualize --open
```

--------------------------------

### Install TWG skills for Claude Code

Source: https://twgcli.com/api

Installs TWG skills for Claude Code, enabling it to use TWG CLI commands.

```bash
# 安裝 TWG skills 給 Claude Code
twg skills install --agent claude

# 或一次裝給所有支援的 coding agent
twg skills install --all-agents

```

--------------------------------

### Explore TWG CLI commands

Source: https://twgcli.com/api

Displays help information for the TWG CLI.

```bash
twg help
```

--------------------------------

### Full JSON interface for a command

Source: https://twgcli.com/api

Returns the precise argument, option, and output schema, including recommendedAgentFields, output presets, and risk warnings.

```json
{
  "type": "cmd",
  "id": "twg:search",
  "cmd": "twg search",
  "args": [{"name": "query", "req": true}],
  "opts": [
    {"long": "--limit", "arg": "<n>"},
    {"long": "--app", "arg": "<app>"},
    ...
  ],
  "output": {
    "recommendedAgentFields": "@compact",
    "primaryCollection": "items"
  }
}

```

--------------------------------

### Fuzzy command search

Source: https://twgcli.com/api

Takes any keyword and returns a relevance-sorted list of commands, useful when only fragments are remembered.

```bash
$ twg help sprint
matches:
  - twg jira sprint
  - twg jira sprint workitems
  - twg confluence search --label sprint

```

--------------------------------

### Top-level command map

Source: https://twgcli.com/api

Returns a concise YAML list of namespaces and their executable subcommands. '$' identifies leaf node commands within a namespace.

```yaml
tree:
  $: [resolve, search, login, logout, help]
  jira:
    board: {}
    sprint: {}
    workitem: {}
  confluence:
    page: {}
    blog: {}
    space: {}
  ...

```

--------------------------------

### Login to TWG CLI

Source: https://twgcli.com/api

Logs the user into the Teamwork Graph CLI.

```bash
twg login
```

--------------------------------

### LangGraph TWG ToolNode

Source: https://twgcli.com/api

Integrates the twg_run function into a ToolNode within LangGraph, enabling nodes in a state graph to query Atlassian work data using the TWG CLI.

```python
from langgraph.prebuilt import ToolNode
from langchain_core.tools import tool
import subprocess, json

@tool
def twg_run(args: list[str]) -> dict:
    """Run a twg CLI command and return parsed JSON."""
    raw = subprocess.check_output(["twg", *args, "-o", "json"])
    return json.loads(raw)

tools = ToolNode([twg_run])

```

--------------------------------

### CrewAI TWG Tool

Source: https://twgcli.com/api

Defines a CrewAI tool that wraps the twg CLI, allowing agents to call any twg command using natural language and receive structured JSON output.

```python
from crewai.tools import tool
import subprocess, json

@tool("twg")
def twg(cmd: str) -> dict:
    out = subprocess.check_output(
        ["twg", *cmd.split(), "-o", "json"]
    )
    return json.loads(out)

```
Tokens:
Raw