Skip to content
Tutorials Center
My learning

Topics

Lesson 30 of 37 ยท Reporting and Response

Writing Up Your Findings: The Analysis Write-Up

Structure a malware analysis write-up for technical and executive readers: summary, sample details, findings, indicators, detections and recommendations, with a runnable generator.

Analysis that nobody can read is analysis that nobody can use. The write-up is the product of your work: the incident responder reads it to scope a breach, the detection engineer reads it to write rules, the manager reads the first paragraph to decide what to do. A good write-up is accurate, short enough to be read, and honest about what you do not know.

This lesson gives a structure you can reuse. The examples use an invented sample called "DemoLoader" and invented indicators. The Python stand-ins turn structured notes into a document and check it for common omissions; real write-ups live in your team's document or case-management system.

Three readers, one document

Decision maker What happened, how bad, what now?
Responder What to search, what to clean?
Analyst How do I verify and extend it?

Answer each reader in turn.

1 / 2

The structure

1. Executive summary (3-6 sentences). State the verdict (malicious, suspicious, benign), the sample's type (loader, backdoor, information stealer), what it does in plain words, the likely impact, how confident you are, and what you recommend. A reader who stops here should still act correctly.

2. Sample details. Names, size, file type, hashes (MD5, SHA-1, SHA-256), fuzzy and import hashes, compile timestamp (with a warning that it can be forged), packer, and where and when you obtained the sample. Include the chain of custody if the case may go to legal review.

3. Methodology and environment. Which tools and which lab (virtual machine snapshot, network simulation, operating system and patch level). This lets others reproduce and lets readers judge how much to trust the results.

4. Static analysis findings. Structure, imports, notable strings, embedded resources, packing and entropy results.

5. Dynamic analysis findings. What happened when it ran: processes, files, registry, network. Present a timeline with relative times.

6. Code analysis (when you did it). Key routines, algorithms (for example the decoding loop and key), configuration format, command set.

7. Capabilities and ATT&CK mapping. A table linking observed behaviours to techniques, covered in the next lesson.

8. Indicators of compromise. Clean lists, defanged, with context (see indicators).

9. Detections. YARA, Sigma and network rules, with a note on testing (tuning).

10. Recommendations. Containment, eradication, hardening, and what to monitor.

11. Limitations and open questions. What you could not do or confirm: the sample needed a missing component, the server was offline, code paths were not reached.

Writing rules that help readers

  • Lead with the conclusion, then support it. Do not make readers wade through chronology to find the verdict.
  • Separate observation from inference. "The sample wrote updater.exe to C:\Users\Public\demo (observed)" versus "this is probably persistence staging (inference)". Mark confidence: confirmed, likely, possible.
  • Show evidence. Screenshot, log excerpt or hash for each important claim. Put long listings in an appendix.
  • Avoid attribution unless you can support it. Code reuse, shared infrastructure and shared language are weak evidence. Say "overlaps with" rather than "is".
  • Use plain words in the summary, precise words in the details. Spell out acronyms once.
  • Defang and label. Use the defanged format for network indicators and mark the TLP (Traffic Light Protocol) sharing level, such as TLP:AMBER, on the first page.
  • Date everything and keep a revision history. Analyses evolve.

From notes to document: a generator

Many analysts keep notes in structured form while they work and generate the text from them, which prevents forgetting sections. The stand-in below builds a Markdown-style document from a dictionary of notes and then runs a checklist over the result.

Python Generate a write-up from structured notes

Press Run (or Ctrl + Enter) to see the result here.

Program input ยท what you'd type while the program runs

Type the answers before you press Run, one per line. Each time the program reads a line (with input()), it gets the next line from this box.

The document separates what was observed from what is only likely, defangs the indicators and ends with limitations. The checklist is intentionally small; teams often add items such as "TLP marking present" and "all hashes are 64 hex characters".

Building a timeline

For the dynamic section, a timeline is clearer than prose. Record the time since execution started and a short, factual description of each event, then sort. The next stand-in merges events from different sources (process monitor, network capture, file system) and prints them in order.

Python Merge event sources into one timeline

Press Run (or Ctrl + Enter) to see the result here.

Program input ยท what you'd type while the program runs

Type the answers before you press Run, one per line. Each time the program reads a line (with input()), it gets the next line from this box.

Gaps in a timeline are information: a 60-second gap between identical requests is a beacon interval; a long quiet period right at the start might be a sleep to outlast a sandbox timeout.

Keep your raw notes, screenshots and logs (the evidence archive) next to the write-up. When a colleague asks "how do you know?", the answer should be one file away.

Check your understanding

Quick check ยท 4 questions

of 4 answered You scored / 4 ยท perfect!

  1. 1. Where should the verdict and impact appear?

    Decision makers may read only the first paragraph, so lead with the conclusion.
  2. 2. Why separate observed facts from inferences?

    Marking observations versus likely or possible conclusions makes confidence transparent.
  3. 3. How should network indicators appear in a shared document?

    Defanging prevents accidental visits and automatic link previews.
  4. 4. What is the purpose of a limitations section?

    Honest limits protect decisions made from the write-up.

Summary

  • A write-up serves three readers: decision makers, responders and fellow analysts.
  • Follow a fixed skeleton: summary, sample details, method, static, dynamic, code, ATT&CK, indicators, detections, recommendations, limitations.
  • Lead with the conclusion, separate observation from inference, show evidence and mark confidence.
  • Defang indicators, add a TLP marking, date the document and keep an evidence archive.
  • Generate write-ups from structured notes and use timelines for dynamic findings.
  • The generator is a stand-in for your team's tools.

Next: Mapping Behaviour to MITRE ATT&CK