| name | jira-worklog |
|---|---|
| description | Logs time on Jira issues and reports logged work. Add worklog entries (with comment and start time), review what you logged today / this week, and roll up time per issue or epic — all via the jira-cli skill. Use when the user says "залогируй время", "log 2h on EDS-1234", "сколько я налогировал", "worklog report", or tracks time against tickets. Works with any Jira project. |
| license | MIT |
| tags | jiratime-trackingproductivity |
| owners | redickowii |
| metadata | {"author":"redickowii","version":"1.0"} |
Jira Worklog
Log and report time on Jira issues. Jira access goes through the jira-cli
skill. Two modes: add (write a worklog entry) and report (read back
logged time).
##Trigger
User invokes /jira-worklog, or asks to log time on / report time for an issue
key (<PROJECT>-<NUMBER>). Works with any project.
##Time format
Jira durations: w weeks, d days, h hours, m minutes — combine with
spaces, e.g. 2h 30m, 1d 4h, 45m. A "day" is the Jira-configured working
day (often 8h), not 24h. If the user gives "полтора часа" / "an hour and a
half", convert to 1h 30m.
##add — log time
Confirm the parsed values (issue, duration, optional comment, optional start time) before writing. Then:
jira issue worklog add <KEY> "<duration>" --no-input \
[--comment "<what was done>"] [--started "<YYYY-MM-DD HH:MM>"]Feed < /dev/null to stop jira-cli hanging on a prompt.
- Default start time = now; only pass
--startedif the user backdates the work. - Never write without confirmation; default is no.
- Verify: re-fetch
jira issue view <KEY> --raw(or worklog list) and confirm the entry landed — jira-cli sometimes prints success without writing.
##report — read logged time
jira-cli has no rich worklog query, so build reports from issue data + JQL.
- Today / this week, mine:
(usecode
jira issue list -q "worklogAuthor = currentUser() AND worklogDate >= startOfWeek()" \ --plain --columns key,summary,timespent --no-headersstartOfDay()for today;worklogAuthor/worklogDateare Jira JQL functions on Server/Cloud). - Per issue:
jira issue view <KEY> --raw→ readtimetracking(timeSpent,originalEstimate,remainingEstimate) and the worklog array. - Per epic: list epic children, then sum
timespentacross them.
Present a compact table: issue · time spent · (estimate / remaining if set). Sum the total at the bottom. State the period and author explicitly.
##Rules
- Convert natural-language durations to Jira format before writing; echo the parsed value back for confirmation.
- Never log time without explicit confirmation; default is no.
- One retry max on a failing jira-cli call, then report the exact error and stop.
- Caveman-mode does NOT apply to worklog comments written into Jira — write them in normal prose.