1
2
{
  "command": "jira workitem query",
  "request": { "site": "example", "jql": "..." },
  "data": {
    "edges": [
      {
        "node": {
          "key": "PROJ-123",
          "summary": "Fix login bug",
          "status": { "name": "In Progress" },
          "webUrl": "https://example.atlassian.net/browse/PROJ-123"
        }
      }
    ]
  }
}

Read the response

Lists — access items via data.edges[].node
Single items — access the result directly via data
Fields are flat — use .summary, .status.name, .webUrl, not .fields.summary

Process JSON output

1
2
# Pipe to jq
twg jira workitem query --jql "project = PROJ" -o json | \
  jq '.data.edges[].node | {key, summary, status: .status.name}'

# Write large results to a file
twg work query --scope me --since 30d --output-file /tmp/work.json

Use agent mode

When called by an agent, append --mode agent. This produces quieter output and strips the hints and warnings keys from the JSON envelope so the agent receives a smaller, cleaner payload:

1
2
twg work query --scope me --since 7d --mode agent

The standard envelope includes:

1
2
{
  "command": "...",
  "request": { ... },
  "data": { ... },
  "hints": ["..."],
  "warnings": ["..."]
}

With --mode agent, hints and warnings are omitted:

1
2
{
  "command": "...",
  "request": { ... },
  "data": { ... }
}

For large result sets, pair --mode agent with --output-file to write JSON to a file and avoid stdout size limits. The CLI prints stdout=<path> to stdout so the agent knows where to read the result:

1
2
twg work query --scope me --since 30d --mode agent --output-file /tmp/work.json
# stdout: stdout=/tmp/work.json