Day 14: Triage Replies, Duplicates and Your Personal Playbook

You submitted a report. Two days later an email arrives with one line: "Thanks, but we cannot reproduce this. Can you provide more information?" Your stomach drops. It is tempting to fire back something defensive, or paste the whole thing into an assistant and send whatever comes out.

Today I will show you how to read that kind of reply for what it really says, answer it in a few calm lines, and then turn everything you have built over this course into one page you can reuse. The reply is a normal part of the process. Handled well, it builds your reputation.

What triage replies actually mean

Triage is the stage where the programme's staff check whether your report is real, in scope and new. Their replies fall into a few common types. Each one needs a different response, so the first skill is naming which one you got.

Notice that only one of these is about your skill. The others are often about timing, policy or impact. A duplicate says you found something real. Treat it as proof your method works, and look at how fast you were.

Triagers read many reports a day. A short, factual, friendly reply gets read. An emotional one gets skimmed, and you want to be the researcher they are glad to hear from.

Watch me do it: answering a request for more information

Here is a mock reply. Imagine you reported a reflected XSS in a lab and the triager wrote the message below. Everything here is practice material on your own machine, so no real programme is involved. I will walk through my method, then you will repeat it.

Hi, thanks for your report. We followed your steps but the alert
did not fire. Which browser did you use, and can you confirm
the exact URL you visited?
  1. Make the folder. In your terminal run mkdir -p day14 and cd day14. Paste the mock message into triage-mock.txt with your text editor. Run ls and you should see the file.
  2. Underline what they asked. They asked two things: the browser, and the exact URL. Write those two items at the top of a new file called reply.md. Answer only what was asked.
  3. Check your own notes for facts. Open your day13 notes.md. Copy the real browser name and version and the exact URL you tested. If a fact is missing, do not guess. Re-run the test and record it.
  4. Draft the reply. Write it in four short parts: thanks, the two answers, one extra helpful detail such as a curl command, and an offer to help further.
  5. Confirm it with curl if relevant. If the URL is a lab on your own machine, run curl -s "URL" | grep -i script and check the payload appears in the response. That evidence goes in the reply.
Hi, thanks for taking a look.

- Browser: Firefox 126 on Linux (private window, no extensions)
- URL: http://localhost:8080/search?q=%3Cscript%3Ealert(1)%3C/script%3E

The payload is reflected unencoded; curl output attached.
Happy to record a short screen capture if useful.

Thanks again.

What just happened: I answered the question in a form the triager can verify in a minute. I added no new claims. If I use an assistant to polish the tone, I give it only the facts from my notes and diff the result, exactly as on Day 13.

Never paste the triager's full message with account details, tokens or private URLs into an assistant. Redact first, as we did on Day 2.

Duplicates, informative and not applicable

Not every reply needs a long answer. Here is how I handle each of the other types.

  1. Duplicate. Thank them, ask nothing more. Record the date you found it and the date you submitted, so you can see how much faster you need to be.
  2. Informative. Read their reasoning. If you think they missed an impact, send one short message with a concrete, proven scenario. If you cannot prove it, accept the decision.
  3. Not applicable. Re-read the policy scope. If you were out of scope, learn from it. If you believe it was in scope, quote the exact policy line once, politely.

Ask for a second look at most once. Pushing repeatedly can get you flagged as difficult and may cost you access to private invitations.

Disclosure rules and staying in scope

After you submit, two rules still bind you. First, disclosure: you do not publish details of a finding unless the programme allows it, and many require written permission. Second, scope: a triage conversation never gives you permission to test further.

If a triager asks for another proof of concept, stay inside the same endpoint and the same test account you used before. Do not look at other users' data, do not escalate to see what else is reachable, and do not retest from a new angle without checking the policy.

Sharing a redacted screenshot of a live programme finding in a chat group is still disclosure. Wait for the programme to say it is fine.

In this course, all practice stays on labs and your own machine, so you can discuss lab results freely. Real programmes follow their own policy, so check it each time.

Build your personal playbook

Over thirteen days you have written prompts, checklists and notes. They are scattered across folders. A playbook is one page you open at the start of every session so you do not rebuild from memory. Keep it short enough to print.

  1. Collect your material. Run ls ../day* and open your best notes. Pick the three prompts that gave you the most useful, verified results, and one checklist you actually used.
  2. Create the file. In your day14 folder create playbook.md with these headings: Before I start, Safe prompting, Testing checklist, Verify before I claim, Report skeleton, Replying to triage.
  3. Fill each heading with two to four lines. Example: under Before I start, write "Read scope. Note what is excluded. Use my own test account."
  4. Add one prompt template. Keep it with placeholders, such as the one below, so you paste facts in instead of rewriting.
  5. Check the length. Run wc -w playbook.md. If it is over about 400 words, cut the least used lines.
Role: careful security reviewer.
Task: tighten the report below. Do not add facts.
Facts I proved: [PASTE]
Facts I did not test: [PASTE]
If something is unclear, ask me instead of guessing.

Reply and playbook

Practise answering triage feedback and consolidate your work. Both pieces are small, and together they become the base of your Day 15 capstone workflow.

  1. Create day14/ and save the mock triage message as triage-mock.txt
  2. Write reply.md answering both questions with facts from your own notes
  3. If you used an assistant to polish it, diff the result and note what you rejected
  4. Write one-line responses for Duplicate, Informative and Not applicable
  5. Create playbook.md with six headings, three prompts and one checklist
  6. Run wc -w playbook.md and trim until it fits one page