Zentao Automatic Timesheet Assistant

Why It Exists

Logging hours in Zentao often means opening a task page, finding the right task, and entering time manually. That workflow is easy to skip when development or project work is in motion. zentao-autoreport targets that gap: when a user says something like “I worked on xxx for x hours today,” the skill matches the description to tasks under the current user, asks for confirmation, and then writes the hours through Zentao’s API.

How It Works

The flow is built around setup, matching, confirmation, and submission. On first use, it asks for ZENTAO_URL, ZENTAO_ACCOUNT, and ZENTAO_PASSWORD, and stores them in /root/.config/zentao/.env. It then fetches the user’s tasks, semantically matches the work description against task names or context, and lists candidates for the user to confirm. If the match is poor, the user can request a rematch or specify a taskID. After confirmation, the skill logs in again to refresh the zentaosid, reads the current remaining hours left, computes newLeft = left - consumed, and submits the result via the recordworkhour endpoint. The scripts prefer the Python implementation and fall back to shell when Python is unavailable.

Boundaries and Caveats

This skill is intended for task time tracking in Zentao 21.x open-source deployments, especially when teams already use Zentao’s task module and want to log hours from conversational notes. It assumes the user’s tasks can be queried, and it still requires human confirmation to avoid writing time to the wrong task. Credentials are stored in a local config file, so they should be managed according to security requirements. If Zentao’s API version, session behavior, or task permissions change, the endpoint path or authentication logic may need adjustment.