Terminal
Terminal/Skill
Helps an agent run on-call handoffs, incident investigations, postmortems, reliability reviews, and risk readouts from Atlassian data.
By Atlassian
Works with
npx skills add atlassian/twg-cli --skill twg-operational-healthUse together with the root twg skill. Exact command grammar comes from live
twg help or twg help describe <path>.
Run twg <command>. On shell command not found, use $HOME/.local/bin/twg
(macOS/Linux) / $env:LOCALAPPDATA\Programs\twg\bin\twg.exe (PowerShell), then
tell user to add that directory to PATH. Do not treat auth or command errors as
PATH failures.
Resolve scope, window, anchors, owner/escalation path, status, recency, and
follow-ups. Find the operational anchor before joining relevant surfaces.
Run without --site; TWG inherits the user's pinned Jira/JSM site. Only add a
site override when the user explicitly requests another tenant. Never assume a
vendor-internal incident site.
working theory, confirmed problem, mitigation, and root cause. Require a causal mechanism for confirmed root cause; a closed record
is insufficient. Mark claims confirmed, supported, candidate, or
missing evidence; use references for RCA details.bitbucket pipeline query, not pipeline get per
pipeline; reserve that for failures whose logs or test reports change the
pattern, owner, or action. query lists recent runs only (--limit default
15), so raise it and state the span covered. Hydrate ranked sets through one
batched get.Load references/reliability-review.md. Resolve leader, platform, and window;
cluster supported themes, separate cause from mitigation, connect prevention
work, and rank leadership actions. For handoffs, add a first-hour checklist and
escalation map.
Use when the incident is active, newly mitigated, or pre-PIR. Anchor on the
incident record, then pull responders, symptoms, impact, recent
deploys/flags/config, topology, alert/log/metric pointers, ownership, runbooks,
and similar incidents. If ticket fields are sparse, probe the four golden-signal
families with bounded follow-ups. See references/incident-investigation.md.
Output a four-signal matrix, hypotheses, confidence, next checks, and mitigation
options; never call a mitigation the root cause without the causal mechanism.
Use after mitigation/recovery when drafting or evaluating a postmortem/PIR. Pair
the incident with the PIR, linked docs, final comms, remediation PRs, and action
items; cover confirmed mitigation, causal mechanism, 5-why chain, and
detection/response gaps. See references/pir-root-cause.md. Output root cause,
contributing factors, mitigation-versus-cause, and prioritized actions.
Build contributors from project/goal/Jira/PR/doc/activity evidence. Inspect
Assets schema/type metadata before AQL; join people via discovered user-like
attributes such as Calculated user. Rank by contribution centrality plus asset
risk, and report confidence and gaps. See references/assets.md.
For staffing, resolve project/topic/org and identify people by related work, ownership, review influence, docs, and project/goal involvement; check load before recommending. For meetings, query scoped recordings, preview transcripts first, fetch full transcripts only for central ones, then summarize decisions, action items, and gaps.
references/assets.md - schema-first Assets queries and person/device joinsreferences/reliability-review.md - bounded incident/PIR leadership reviewreferences/incident-investigation.md - active investigation and mitigationreferences/pir-root-cause.md - post-mitigation root-cause and PIR workflow