Sprint Snapshot

Sprint Snapshot

Point-in-time snapshot of a scrum team's current sprint. Read-only across Jira. Writes three artifacts per run into the team's vault folder so progress can be tracked over time, scrum-of-scrums and end-of-sprint reporting become low-friction, and trends can be forecast.

This skill is a Claude Code port of <scripts-repo>\Sprint-Planner.ps1 + <scripts-repo>\Get-Jira-Sprint-Planner.ps1 — same canvas layout, same output path convention, same JQL shape. See REFERENCE.md for the canvas layout spec, JSONL schema, and matching algorithm. See EXAMPLES.md for invocation parity with the PS scripts.

Quick start

/sprint-snapshot                                  → Aurora, INC 28, current sprint, auto phase
/sprint-snapshot Borealis                             → single team override
/sprint-snapshot Aurora --sprint 2 --inc 28       → explicit sprint within increment
/sprint-snapshot Aurora --as-of 2026-04-08        → historical snapshot (Jira `updated <= ...`)
/sprint-snapshot Aurora --phase "week 1"          → override the auto-derived phase

Parameters

JIRA project key, vault root, GitLab base, default team, capacity, and velocity are all resolved from memory and the per-sprint config note — never accept them as raw args. See Config resolution below.

Config resolution (run once at the start of every invocation)

If a memory file is missing or its value is <unset>, prompt once via AskUserQuestion with the PS-script default as the recommended option, then write the answer back so the next run skips the prompt. Same pattern as daily-standup-prep.

Sprint config note (per-sprint frontmatter)

{{sprint_config}} is the durable source of truth for capacity/velocity per sprint. Schema:

---
team: Aurora
increment: 28
sprint: 2
start_date: 2026-04-01
end_date: 2026-04-21
capacity: 80.8
last_sprint_velocity: 55
avg_velocity_last_3: 58
labels: [pyrite, dev_common, dev_feature, doc_feature]
---
# Sprint 2 — Aurora (INC 28)
Notes, links, retro pointers go below.

If the file doesn't exist when the skill runs, prompt for start_date, end_date, capacity, last_sprint_velocity, avg_velocity_last_3 via a single AskUserQuestion (recommend the PS-script defaults: capacity 80.8, last_velocity 55, avg 0), then write it. Subsequent snapshots in the same sprint reuse it.

Roster bootstrap (first-run only)

Same pattern as daily-standup-prep. Look for {{rosters_dir}}\<Team>.csv. If missing, offer to copy from <scripts-repo>\bin\Teams\<Team>.csv via AskUserQuestion. CSV format and identity-matching cascade are identical — see daily-standup-prep/REFERENCE.md.

Workflow

- [ ] Phase 1: Resolve config, load roster, load/create sprint config
- [ ] Phase 2: Auto-derive phase + build effective JQL (with optional AsOf)
- [ ] Phase 3: Fetch Jira issues + per-issue details
- [ ] Phase 4: Match assignees to roster + bucket by status
- [ ] Phase 5: Render canvas (delegate to obsidian-canvas)
- [ ] Phase 6: Render companion markdown summary
- [ ] Phase 7: Append snapshot row to _snapshots.jsonl (trend log)
- [ ] Phase 8: Console summary

Phase 1 — Config + roster + sprint config

Resolve every {{placeholder}} per the table above. Load the roster CSV using the cascade in daily-standup-prep/REFERENCE.md. Read or create {{sprint_config}} and parse its frontmatter into {Team, Inc, Sprint, StartDate, EndDate, Capacity, LastSprintVelocity, AvgVelocityLast3, Labels[]}.

Phase 2 — Phase derivation + JQL

Auto-derive Phase (unless user overrode):

State the chosen phase in one short sentence: "Snapshotting Aurora Sprint 2 (INC 28) — phase: week 1 (day 5 of 21)."

Build effective JQL (port of <scripts-repo>\Get-Jira-Sprint-Planner.ps1 lines 24, 554-570):

project = {{jira_project}}
  AND labels in ({comma-joined Labels[] from sprint config})
  AND Sprint in openSprints()
  AND Sprint not in futureSprints()

If --as-of <date> was supplied, append AND updated <= "{YYYY-MM-DD}" and parse via standard date handling. Echo the final JQL one line before issuing the search so the user can confirm.

Phase 3 — Fetch Jira issues

Single call:

jira_search(
  jql=<effective JQL>,
  fields="summary,description,status,assignee,priority,issuetype,labels,customfield_10106",
  max_results=200
)

customfield_10106 is Story Points (matches the PS script). Don't refetch per-issue unless the search response is truncated. Cap at 200 — warn if hit.

Phase 4 — Match + bucket

For each issue:

1. Parse story points: treat as double; missing → 0. Track total + per-status running totals.
2. If assignee present, run the identity cascade against the roster. Match → team_member. No match → off_team_member (still tracked, separate group on canvas).
3. Bucket the issue by status:
   • In Progress → IN PROGRESS column
   • In Review → IN REVIEW column
   • Done / Won't Fix / Duplicate → DONE column
   • everything else → TO DO (split into Assigned vs Unassigned sub-columns based on whether assignee is set)
4. Increment per-member workload (TicketCount, StoryPoints) for matched team and off-team members.

Compute summary stats: totalTickets, totalStoryPoints, per-status counts and points, unassignedTickets/Points, ticketsWithPoints, ticketsWithoutPoints, percentDone (tickets), percentPointsDone.

Phase 5 — Render the canvas

Delegate to obsidian-canvas via Skill for actual JSON authoring, ID generation, and validation — do not hand-author canvas JSON inline. The canvas layout (groups, node positions, sizes, colors, edges) is fully specified in REFERENCE.md. Pass the rendered structure to obsidian-canvas.

Issue card content uses the JIRA:KEY convention so each card auto-links to the live ticket via the user's Obsidian Jira plugin:

## [PROJ-1234] Summary line truncated to ~80 chars

JIRA:PROJ-1234

**Assignee:** Alex Rivera
**Points:** 3
**Type:** Story

The JIRA:KEY line goes on its own line — the plugin auto-renders. Do not wrap it in a wikilink. Team-member cards use the wikilink form [[@First Last]] (verify the vault note exists via Glob("{{vault_root}}/🤼 Team/**/@*.md"); if missing, fall back to bold plain text — same rule as daily-standup-prep).

Output path: {{output_root}}\<phase>.canvas — e.g. …\Sprint 2\week 1.canvas. Phase label is preserved literally including the space.

Never overwrite silently. If <phase>.canvas exists, prompt: Overwrite / Save with timestamp suffix (week 1 (HH-MM).canvas) / Skip canvas. Same pattern as daily-standup-prep.

Phase 6 — Companion markdown summary

Write {{output_root}}\<phase>.md next to the canvas. Schema is the Sprint Summary table from the PS script (lines 973-996) plus a small header. Delegate to obsidian-markdown if extending with callouts, dataview, or new constructs. Template:

---
team: {{team}}
increment: {{inc}}
sprint: {{sprint}}
phase: {{phase}}
snapshot_date: {{YYYY-MM-DD HH:mm}}
as_of: {{as_of_or_blank}}
total_tickets: {{n}}
total_story_points: {{n}}
percent_done_tickets: {{n}}
percent_done_points: {{n}}
canvas: "[[{{phase}}.canvas]]"
---
# Sprint {{sprint}} — {{phase}} snapshot

> Captured {{YYYY-MM-DD HH:mm}} (Pittsburgh){{ as_of? " — as-of " + as_of : "" }}

## Sprint Summary

| Metric | Value |
| :--- | --: |
| Total Tickets | … |
| Total Story Points | … |
| Tickets Done / Open / In Review / In Progress / To Do | … |
| Unassigned Tickets / Points | … |
| % Tickets Done / % Points Done | … |
| Tickets With Points / Without Points | … |
| Points Done / In Review / In Progress / To Do | … |
| Capacity / Committed / Remaining | … |
| Last sprint velocity / 3-sprint avg | … |

## Per-member workload

| Member | Tickets | Points | % of capacity |
| :--- | --: | --: | --: |
| [[@First Last]] | … | … | … |

## Off-team workload

(only rendered if non-empty)

## Velocity trend

(↗️ Increasing / ↘️ Decreasing / → Stable — port of PS lines 305-317)

Same overwrite-prompt rule as Phase 5.

Phase 7 — Append to trend log

Append a single JSON line to {{output_root}}\_snapshots.jsonl. Create the file if missing — never overwrite. Schema is in REFERENCE.md. Minimum fields:

{"snapshot_at":"2026-04-08T09:14:00-04:00","as_of":null,"phase":"week 1","team":"Aurora","inc":28,"sprint":2,"capacity":80.8,"committed":62.5,"remaining":18.3,"totals":{"tickets":24,"points":62.5},"by_status":{"todo":{"tickets":8,"points":21},"in_progress":{"tickets":6,"points":15.5},"in_review":{"tickets":4,"points":11},"done":{"tickets":6,"points":15}},"unassigned":{"tickets":3,"points":8},"velocity":{"last":55,"avg3":58}}

This append-only log is what enables forecasting and trend analysis without parsing canvas files. Keep one row per snapshot — never rewrite history.

Phase 8 — Console summary

Print a brief block (matches the PS Write-Verbose summary, lines 893-903):

=== Aurora Sprint 2 (INC 28) snapshot — phase: week 1 ===
JQL:        project = PROJ AND labels in (pyrite,dev_common,...) AND Sprint in openSprints()...
Issues:     24 total / 6 done (25%) / 4 in review / 6 in progress / 8 to do
Points:     62.5 total / 15 done / 11 in review / 15.5 in progress / 21 to do
Unassigned: 3 tickets (8 pts)
Capacity:   80.8 (committed 62.5, remaining 18.3)
Saved:      [week 1.canvas](Scrum Teams/Aurora/Scrum 📅/INC 28/Sprint 2/week 1.canvas)
            [week 1.md](Scrum Teams/Aurora/Scrum 📅/INC 28/Sprint 2/week 1.md)
            [_snapshots.jsonl] (+1 row, 4 total)

Rules

• Read-only across Jira. Never transition issues, post comments, change assignees, or touch the API beyond jira_search / jira_get_issue.
• Append-only on _snapshots.jsonl. Never rewrite or compact. The whole point is point-in-time history.
• Never overwrite an existing canvas or markdown summary silently. Always confirm via AskUserQuestion.
• Confirm before persisting to memory or to the sprint config note. When populating a missing value, use the PS-script default as the recommended option but let the user override before writing.
• Phase label is preserved literally including the space (week 1, not week-1) — the PS script's filenames are the canonical convention and other tools may consume them.
• Pittsburgh local time for snapshot timestamps and phase derivation. Jira updated cutoffs use ISO date (YYYY-MM-DD).
• No fabricated data. If Jira fails, surface the failure in the console summary and abort. Do not invent issues. (Unlike daily-standup-prep, partial outputs aren't useful here — a snapshot must reflect real Jira state.)
• Wikilinks for team members only when the vault person note exists. Use Glob("{{vault_root}}/🤼 Team/**/@*.md") to verify before writing [[@First Last]]. Fall back to bold plain text otherwise — never invent a wikilink.
• JIRA:KEY for issue references. This is the Obsidian Jira plugin's auto-link convention — do not wrap in a wikilink, do not add the URL inline.
• Story Points = customfield_10106. This is the PS-script's hardcoded ID and it matches the user's Jira instance. If the field is missing for an issue, treat as 0 and increment ticketsWithoutPoints.
• Auto-derived phase is a suggestion. Always echo the derived phase in Phase 2 output so the user can rerun with --phase if they disagree.

Edge cases

• Sprint config note missing or incomplete — prompt once for the missing fields, write the file, continue. Don't run with synthetic capacity/velocity.
• Roster CSV missing — offer to copy from the upstream PowerShell scripts repo (same pattern as daily-standup-prep); if neither path exists, abort the team.
• labels in (...) returns 0 issues — render an "empty sprint" canvas (groups + columns, no issue cards), still write the markdown summary and JSONL row. Surface the empty result clearly so the user can fix the label set.
• AsOf date is in the future — reject with an AskUserQuestion: did you mean today, or a different date? Don't silently fall through to "current".
• Issue without an assignee — goes in TO DO (Unassigned) column; counted in unassignedTickets/Points.
• Assignee not on roster — goes in the off-team group on the right of the canvas; tracked separately in offTeamMembersWorkload. Don't pollute the team workload total with off-team points.
• Vault 🤼 Team folder missing the person note — render bold plain text instead of a broken wikilink. Same rule as daily-standup-prep.
• Re-run on the same phase — overwrite/timestamp/skip prompt for .canvas and .md. The JSONL row is always appended (it's the point-in-time history).
• customfield_10106 field name differs in the user's Jira — the field ID is the user's confirmed Story Points field per the PS script. If a future Jira config change breaks it, the skill will count everything as 0 points and surface the anomaly in the console summary; the fix is to update the field ID in reference_jira_default_project.md (add a **Story Points field:** line).
• Pre-sprint snapshot (AsOf = sprint start) — the PS script's intended use case for start phase. JQL filter still works because the issues already have the sprint label by then.

Related skills

This skill writes into the user's Obsidian vault and reads from Jira. Delegate to the relevant sibling skills via Skill rather than reimplementing.