Skip to content

rampstackco/claude-skills / skills/backlog-triage

backlog-triage

Verified

Turn an inbox of raw requests (bug reports, feature asks, support escalations, feedback, incident follow-ups) into a ranked, deduplicated backlog where every item at the top meets the definition of ready, with recorded scores, merge links, and a written reason for every park and decline. Use this skill whenever the user wants to triage incoming work, groom or prune a backlog, merge duplicate tickets, score or rank requests, decide what the team works on next, or run a weekly triage session. Triggers on triage, groom the backlog, backlog grooming, prioritize the backlog, dedupe tickets, duplicate tickets, what should we work on next, inbox of bugs, feature requests pile, rank these, is this ready, backlog is too big, stale tickets, severity vs frequency. Also triggers when the user pastes a list of raw requests or bug reports and asks what to do with them, even if they never say the word triage.

references/scoring-template.md

View on GitHub
# Scoring Template

Two scoring sheets: one for bugs, one for stories and epics. Both record inputs on the item so the score can be re-derived by someone who was not in the room.

The rule for both: if an input is unknown, the item is "unscored" with the missing input named. Do not fill the gap with a plausible number.

---

## Bugs: severity and frequency

Score severity and frequency as two separate fields. Do not combine them into one number. The grid decides queue position.

### Sheet

```
ITEM: [ID and title]
SEVERITY: [S1 / S2 / S3 / S4]  (scale from the ticket-authoring bug template)
Severity evidence: [what makes it this level, e.g., "payroll totals wrong, no workaround"]

FREQUENCY: [high / low]
Frequency evidence:
- Reporter count: [n]
- Affected accounts or users from logs or analytics: [n, source, date range]
- Reproduction rate: [every time / N of M / once]

GRID CELL: [high sev + high freq / high sev + low freq / low sev + high freq / low sev + low freq]
QUEUE POSITION: [from the grid below]
Override (if any): [reason, e.g., "contractual SLA for this account"]
```

### The grid

| | Low frequency | High frequency |
|---|---|---|
| **High severity (S1, S2)** | Next sprint. Note a workaround now. | Top of the queue. Consider interrupting the current sprint. |
| **Low severity (S3, S4)** | Park or decline. Revisit if frequency rises. | Batch with nearby work. Queue behind high-severity items. |

Set the frequency line for your product before you use the grid. For example: "high frequency" means more than 1 percent of weekly active accounts, or more than 5 reporters in 30 days. Write the line down so the grid gives the same answer each week.

---

## Stories and epics: value against effort

Use the model your team already uses. If there is none, RICE is one option. Record every input and where it came from.

### Sheet (RICE shown)

```
ITEM: [ID and title]

REACH: [number of users or accounts per period]
Source: [analytics query, reporter count, segment size]

IMPACT: [3 massive / 2 high / 1 medium / 0.5 low / 0.25 minimal]
Rationale: [one line on what changes for the user]

CONFIDENCE: [100% / 80% / 50%]
Rationale: [what evidence exists: research, reporter count, experiment, hunch]

EFFORT: [person-weeks]
Source: [engineer estimate, name and date]

SCORE: (Reach x Impact x Confidence) / Effort = [n]
```

### Worked line

Reach 400 accounts per quarter (from the usage report), impact 1 (saves a weekly manual step), confidence 80 percent (11 reporters, no research), effort 2 person-weeks (engineer estimate). Score: 400 x 1 x 0.8 / 2 = 160.

### Other models

- **Value, urgency, effort:** three 1 to 5 ratings with a one-line rationale each. Score: (value + urgency) / effort.
- **Cost of delay over duration:** estimate what a week of delay costs, divide by weeks of work. Strong when cost of delay is measurable.

Pick one per team and keep it for at least a quarter. Switching models every month makes every ranking incomparable to the last.

---

## Tasks and spikes

Tasks inherit the score of their parent story. Spikes inherit the score of the decision they unblock. Record the parent, not a new score.

---

## Ties and overrides

- **Ties:** break by cost of delay, then by age (oldest first).
- **Overrides:** allowed, written on the item with a reason and a name. "Customer contract requires a fix by the 30th, approved by [name]." An override with no name is loudest-voice ordering.
reverse-engineer-anythingmorluto/reaReverse engineer native, managed, Electron/JavaScript, packaged, firmware, and browser targets with REA, and build agent environments (AgentEnv bundles) of web applications from a live instance or a capture, and of Linux desktop apps with a task from a request. Use shipped-artifact or requested runtime evidence to explain features, compare versions, decompile code, or guide a reconstruction. Skip REA for ordinary source-repository architecture analysis.Stars: 471kUpdated Claude CodeCodexVerifiedInstall
diagnosing-superpowersobra/superpowersUse when a superpowers session went wrong and your human partner wants to know why — repeated work, ignored plans, stumbles, poor results, a skill that didn't fire, "it took too long", "why is it so expensive", "what is it doing" — or wants to build a bug report for the superpowers maintainers, for the current session or a past one identified by id or path, on any harness.Stars: 322kUpdated Claude CodeCodexVerifiedInstall
dispatching-parallel-agentsobra/superpowersUse when facing 2+ independent tasks that can be worked on without shared state or sequential dependenciesStars: 322kUpdated Claude CodeCodexVerifiedInstall
executing-plansobra/superpowersUse when executing an implementation plan in the current session as the implementer yourself — your human partner chose inline execution, or no subagent tool is availableStars: 322kUpdated Claude CodeCodexVerifiedInstall