スキル一覧に戻る
Lordsisodia

feedback-triage

by Lordsisodia

SISO Agency Internal Platform - Task management and day tracking system with LifeLock integration

0🍴 0📅 2026年1月19日
GitHubで見るManusで実行

SKILL.md


name: feedback-triage category: knowledge-documentation/documentation version: 1.0.0 description: Turn messy qualitative feedback into a prioritized, actionable backlog author: obra/superpowers verified: true tags: [feedback, triage, prioritization, backlog, ux, documentation]

Skill: Feedback Triage

This skill is designed for when you have:

  • User feedback
  • UI testing notes
  • Bug reports
  • "It feels confusing" comments

...and you want a stable output that teams/agents can execute against.

<inputs_to_collect> Source(s) of feedback (paths/links) Product area scope (what's in / out) Target user persona (internal ops vs merchants vs customers) Success criteria ("done" definition) </inputs_to_collect>

<common_themes>

  • Navigation & discoverability
  • Data model / simplicity
  • Performance & reliability
  • Feature completeness
  • Docs clarity
  • Onboarding / first-run experience
  • Error handling / edge cases </common_themes>

<prioritization_factors> How many users affected? How severe is the pain? How often does this issue occur? How much work to fix? (quick wins first) What's the risk of not fixing? (security, data loss, etc.) </prioritization_factors>

Priority Score = (Impact + Frequency + Risk) / Effort </scoring_method>

<backlog_item_structure> For each backlog item include: Short, descriptive title (one line) Clear description of the problem (what, where, when) Source link/path, frequency count, user quotes High-level approach to fixing the issue Specific, testable conditions for "done" Which agent/team should handle this Numeric score from prioritization phase Tier 1 / 2 / 3 </backlog_item_structure>

Problem: [Problem statement]

Evidence: [Source link/path]

  • Mentioned [N] times
  • User quote: "[...]"

Proposed Solution: [High-level approach]

Acceptance Criteria:

  • [Specific, testable condition 1]
  • [Specific, testable condition 2]
  • [Specific, testable condition 3]

Suggested Owner: [Agent/team] Priority: Tier [X] (Score: [Y]) Type: [bug/UX/feature/etc] Surface: [UI/backend/database/etc] Estimated Effort: [quick fix/refactor/etc]

Date: [Date] Sources: [List of sources] Scope: [Product area] Persona: [Target user] Total Issues: [N]

Summary

Quick Wins (Do First)

  • [Quick win 1] ([Tier X])
  • [Quick win 2] ([Tier X])

Top Priority Themes

  1. Theme 1 ([N] issues, [X] Tier 1)
  2. Theme 2 ([N] issues, [X] Tier 1)
  3. [Theme 3] ([N] issues, [X] Tier 1)

Themes & Backlog

[Theme Name]

[Theme description - what area this covers]

Issues: [N] total ([X] Tier 1, [Y] Tier 2, [Z] Tier 3) Suggested Owner: [Agent/team]

Backlog Items

[Insert backlog items here]

Statistics

By Type:

  • Bugs: [N]
  • UX Issues: [N]
  • Missing Features: [N]
  • Performance: [N]
  • Data Integrity: [N]
  • Docs Gaps: [N]

By Severity:

  • Blockers: [N]
  • High: [N]
  • Medium: [N]
  • Low: [N]

By Surface:

  • UI: [N]
  • Backend: [N]
  • Database: [N]
  • Architecture: [N]
  • Docs: [N]

Execution Plans

Next Steps

  1. Review and prioritize quick wins
  2. Assign owners to each theme
  3. Create execution plans for Tier 1 items
  4. Schedule review for Tier 2/3 items

<normalization_rules> Convert every vague complaint into an observable problem statement Preserve user quotes in evidence - they carry context If you can't reproduce an issue, mark confidence as "speculative" </normalization_rules>

<best_practices> Always preserve the original feedback in evidence links Group by theme, not by source - users don't think in "docs vs Slack" Quick wins build momentum - flag them prominently If a theme has >10 issues, consider breaking it into sub-themes </best_practices>

Issues: 8 total (3 Tier 1, 2 Tier 2, 3 Tier 3) Suggested Owner: ui-ux-agent

Backlog Items

Add breadcrumbs to deep pages

Problem: Users land on deep pages from search/links and can't navigate back to parent sections

Evidence: docs/feedback/ui-review.md

  • Mentioned 5 times
  • User quote: "I got lost, couldn't find my way back"

Proposed Solution: Implement breadcrumb component showing page hierarchy

Acceptance Criteria:

  • Breadcrumb appears on all pages depth > 2
  • Each crumb is clickable and navigates to correct page
  • Current page is shown as last crumb (not linked)
  • Breadcrumb is visible on mobile (responsive)

Suggested Owner: ui-ux-agent Priority: Tier 1 (Score: 12.5) Type: UX confusion Surface: UI Estimated Effort: quick fix

Issues: 3 total (2 Tier 1, 1 Tier 2) Suggested Owner: backend-team

Backlog Items

Fix duplicate record creation on concurrent submits

Problem: When users double-click submit button, duplicate records are created

Evidence: Slack #bugs, GitHub issue #234

  • Mentioned 7 times
  • User quote: "Created 2 identical orders, now what?"

Proposed Solution: Add idempotency key to submit handler

Acceptance Criteria:

  • Submit handler generates unique idempotency key
  • Backend validates idempotency key before creating record
  • Duplicate submit returns existing record, not error
  • Button disabled after first click

Suggested Owner: backend-team Priority: Tier 1 (Score: 15.0) Type: bug Surface: backend + UI Estimated Effort: refactor

<output_format> <triage_report_structure>

  1. Summary (quick wins, top themes)
  2. Themes & Backlog (grouped by theme, prioritized)
  3. Statistics (by type, severity, surface)
  4. Execution Plans (links to follow-up plans)
  5. Next Steps (actionable items) </triage_report_structure> </output_format>

スコア

総合スコア

50/100

リポジトリの品質指標に基づく評価

SKILL.md

SKILL.mdファイルが含まれている

+20
LICENSE

ライセンスが設定されている

0/10
説明文

100文字以上の説明がある

0/10
人気

GitHub Stars 100以上

0/15
最近の活動

3ヶ月以内に更新がある

0/10
フォーク

10回以上フォークされている

0/5
Issue管理

オープンIssueが50未満

+5
言語

プログラミング言語が設定されている

+5
タグ

1つ以上のタグが設定されている

0/5

レビュー

💬

レビュー機能は近日公開予定です