Sort the pile by harm, reach, and effort. Not by who shouted.
A triage pass that turns a long bug list into a ranked order with a decision for each bug, using only the evidence you supply.
Triage often follows volume. The bug an executive mentioned twice beats a quiet data-loss bug nobody has reported yet. Severity, reach, and effort get blended into a feeling, and the order cannot be explained afterward.
Bug Triage rates the three separately, cites the report text behind each severity, and marks reach and effort as unverified when you gave no numbers. It then ranks, explains every move off the plain order, assigns a decision to each bug, and challenges the top of the list.
What it produces.
- →A normalized list with duplicates merged
- →Severity, reach, and effort per bug, with unverified marked
- →A ranked order with a one-line reason for each exception
- →A decision per bug: fix now, this cycle, backlog, needs info, or will not fix
- ✦Product and engineering leads running a triage meeting
- ✦Support leads who need to explain why a bug is waiting
- ✦Small teams cleaning a backlog before a release
- →bug-triage.zip: a native Claude Skill. Upload to Settings → Skills in Claude, or unzip to
~/.claude/skills/ - →bug-triage-skill.md: plain markdown. Works in ChatGPT Custom GPTs, Gemini, Claude Projects, agent system prompts, anywhere
- →A README inside the zip with install steps for Claude, ChatGPT, and Gemini
What it does with a real question.
I'm the QA lead for our scheduling app, and we ship v4.2 on March 14. I've exported 63 open bugs from Jira into a CSV, with reporter, affected platform, and customer ticket counts for each. Our team has three engineers and about 40 hours of fix capacity before the freeze, so please sort them into a defensible order.
Start with severity, not ticket volume. Your CSV has ticket counts, which is useful, but counts measure how loudly a bug is reported, not how much damage it does. A data-loss or security bug with two tickets outranks a cosmetic glitch with thirty. So my position is this: rate all 63 by worst plausible outcome first, then use ticket counts and platform to break ties within each tier, then use effort to decide what actually fits.
With 40 hours and three engineers, you are not fixing 63 bugs. Realistically you are fixing somewhere between eight and fifteen, so the real work is deciding the cut line and defending it. Anything in the top severity tier goes in regardless of cost, because shipping a known double-booking or data-loss bug on March 14 is the thing you will be asked about afterward. After that, favor the cheap fixes that clear high-reach annoyances: a two-hour fix with fifteen tickets behind it beats a three-day fix with twelve.
Watch for two traps. Duplicates will inflate your counts, so merge them before ranking. And any bug without an engineer's effort estimate cannot be ranked honestly, so label it unknown and do not guess.
Your concrete next move: spend your first hour tagging every bug with one of four severity levels, then hand just the top two tiers, probably 15 to 20 bugs, to your engineers for hour estimates. Ranking the whole 63 before that is wasted effort.
The full Playbook runs the whole procedure and writes the complete deliverable, not this excerpt.
Written by claude-sonnet-5-5 with this Playbook loaded, 2026-10-11. Checked automatically for length and for copied framework text.
No subscription · free updates within v1.x
Works as a Claude Skill (.zip) and as a plain .md for ChatGPT, Gemini & every LLM
Delivered to your email within 60 seconds · Secured by Stripe
Price includes UK VAT where applicable
Before you buy.
How is this different from just asking Claude directly?
Asked cold, Claude will rank bugs confidently from titles and fill in reach it cannot know. This keeps the three questions apart, labels guesses as guesses, and asks for the evidence that would confirm the bugs it ranked highest.
How is this different from a Tool?
Playbooks are systematically produced to a fixed template and priced accordingly ($12.99). Tools are hand-built, one at a time, and go deeper ($39 to $199). A Playbook gives you a reliable method for routine work; a Tool is for when the judgment matters more than the format.
Will this work with ChatGPT, Gemini, and other LLMs?
Yes. You get a native Claude Skill (.zip) that auto-invokes in Claude, plus a plain .md you can paste into a ChatGPT Custom GPT, a Gemini system prompt, or any LLM. The procedure travels with the file.
Can I share this with my team?
It's licensed for your own personal and professional use. The output it produces is yours to circulate. To put the Playbook itself in the hands of a whole team, email hello@authority.md about team licensing.
What if it doesn't work for me?
If the file hasn't been delivered, or arrives defective or empty, we refund in full: see our refund policy.