IR Documentation and Case Management
Good documentation during an incident protects the analysts who worked it, enables the organization to prove it acted reasonably, and converts the incident into institutional knowledge. Bad documentation is discoverable evidence of negligence.
Six months after a significant breach, outside counsel sends a litigation hold notice: all documents and communications related to the incident are subject to discovery. The IR lead starts collecting materials. The timeline exists — sort of — in a shared Google Doc, but it wasn't updated in real time and multiple entries have conflicting timestamps. The decision to delay containment for two hours — a defensible call at the time — appears nowhere in writing. The CISO's Slack message from that night says "let's wait and see if this is real before we go crazy" which, taken out of context, reads like negligent inaction. The forensic images were collected but the chain of custody log was never created, making them potentially inadmissible. The final incident report was never written because the team was exhausted and immediately moved on to hardening work.
The technical work done during this incident was excellent. But without documentation, the organization cannot prove that, and the lawyers are not technical.
Why Documentation Is Not Bureaucracy
IR documentation serves three distinct purposes that have nothing to do with paperwork for its own sake. Understanding these purposes explains why cutting corners on documentation is a security failure, not an efficiency win.
Three Purposes of IR Documentation
══════════════════════════════════════════════════════════════════════
1. LEGAL PROTECTION
─────────────────
The incident will likely be reviewed by:
• Outside counsel in litigation or regulatory investigation
• Insurance adjusters evaluating your claim
• Regulators assessing whether notification was timely
• Your own board in a post-incident governance review
Without documentation, you cannot prove:
• What you knew and when you knew it
• What decisions were made and why
• That actions were taken promptly and reasonably
• That the breach scope is what you say it is
2. INSTITUTIONAL MEMORY
─────────────────────
IR team members leave. Incidents recur months later. Without
documentation, the next IR team:
• Repeats mistakes from the last incident
• Cannot correlate the new incident to the old one
• Has no baseline for "what did our network look like before?"
• Cannot run a meaningful post-incident review
3. PROGRAM IMPROVEMENT
─────────────────────
The only way to improve an IR program is to measure it.
Documentation enables:
• MTTD/MTTR calculation
• Detection gap identification
• Playbook gap identification
• Lessons learned that are specific and actionable
(not "we should communicate better" but "we discovered
at T+6 hours that the asset inventory was out of date
and this delayed scoping by 4 hours — recommendation:
quarterly asset inventory validation")
The Incident Case File
Every incident should have a structured case file that contains all artifacts, findings, decisions, and communications from the incident in one place. This is not a massive overhead — it is a folder structure you create at the start of the incident and populate as you go.
Incident Case File Structure ══════════════════════════════════════════════════════════════════════ IR-2026-023/ ← Incident ID: IR-[year]-[sequence] ├── 00_summary.md ← Incident summary, severity, status │ (updated continuously during IR) ├── 01_timeline.md ← Chronological event log (real-time) ├── 02_scope.md ← Confirmed affected systems/accounts │ (updated as scope expands) ├── 03_decisions.md ← Decision log with full context ├── 04_evidence/ │ ├── evidence_inventory.csv ← Hash, filename, collection time, │ │ collector, storage location │ ├── memory_images/ ← .raw files with .sha256 sidecar │ ├── disk_images/ ← Forensic images │ ├── log_exports/ ← .evtx files, CSV exports, SIEM saves │ └── malware_samples/ ← With hash verification ├── 05_communications/ │ ├── sitreps/ ← All SITREPs (timestamped) │ └── external/ ← Communications with regulators, │ LE, insurance, IR firm ├── 06_tasks.md ← Current task list, owners, status └── 07_pir.md ← Post-incident review (filled after close) Version control: Put this in a Git repository. Access control: Restrict to IR team + legal. Not in the same email system that may be compromised. Backup: Out-of-band storage — not on corporate infrastructure that may be affected by the incident.
If the incident involves email compromise or a broad corporate network intrusion, using corporate email or Slack as the primary communication and documentation channel has two serious problems. First, the attacker may be monitoring those channels — you are documenting your IR strategy for the attacker to read in real time. Second, those communications are the incident record that legal will review. Informal Slack messages read very differently in litigation than they feel in the moment. The rule is: at the start of any significant IR, switch to out-of-band communication for operational discussions (Signal, phone, a dedicated war room), and maintain documentation in a dedicated case file system that is separate from the potentially-compromised corporate environment.
The Incident Timeline: Writing It in Real Time
The timeline is the most important document in any incident. It is the reconstructable record of what happened, in what order, with what evidence. A timeline written in real time — entry by entry as findings emerge — is far more reliable than one reconstructed from memory after the incident closes.
Timeline Entry Format and Examples
══════════════════════════════════════════════════════════════════════
Required fields per entry:
[UTC timestamp] | [Category] | [Finding] | [Evidence] | [Analyst]
Categories: DETECTION, FINDING, DECISION, CONTAINMENT, ERADICATION,
COMMUNICATION, EVIDENCE-COLLECTION, EXTERNAL
Examples:
─────────────────────────────────────────────────────────────────────
2026-09-15 03:12 UTC | DETECTION | EDR alert: 400+ file encryption
events on LAPTOP-JSMITH, LockBit signature. SOC on-call paged.
| EDR Alert ID: CS-2026-0915-8a9f | analyst-taylor
2026-09-15 03:18 UTC | DECISION | Decision: delay containment
20 minutes to scope additional hosts before isolating. Active
encryption appears stopped. Risk accepted: attacker may use window
to expand. Authorized by: IR Lead Martinez. CISO Chen notified via
Signal. | Signal message screenshot (retained in 05_communications)
| ir-lead-martinez
2026-09-15 03:22 UTC | FINDING | EDR fleet sweep identifies 6
additional hosts with beacon activity (list: LAPTOP-JONES,
SERVER-APP02, LAPTOP-PATEL...). LockBit confirmed on all by .lockbit
extension. | EDR host group search ID: SWEEP-20260915-01 | analyst-kim
2026-09-15 04:47 UTC | EVIDENCE-COLLECTION | Memory image collected
from LAPTOP-JSMITH. SHA256: a3b4... Size: 16GB. Collected by
analyst-taylor using WinPmem. Stored: /evidence/memory_images/
laptop-jsmith-20260915.raw. Chain of custody: custody-form-001.pdf
| analyst-taylor
─────────────────────────────────────────────────────────────────────
Key discipline: every analyst who finds something adds an entry
immediately. The timeline is append-only. Never edit previous entries.
The Decision Log
Every material decision made during the incident — especially any decision to delay action or accept risk — must be logged separately from the timeline. The decision log is your most important legal protection. It transforms "we were negligent in not containing immediately" into "we made a documented, authorized, risk-informed decision to delay containment by 20 minutes for specific reasons, authorized by the IR lead and documented at the time."
| Field | What to write | Why this field exists |
|---|---|---|
| Decision ID | DECISION-001, DECISION-002 — sequential | Reference in timeline, communications, and legal analysis |
| Time (UTC) | Exact time the decision was made | Establishes timeline context — "what did you know and when did you know it" |
| Decision | Concrete, specific — not "we decided to contain" but "we decided to delay full network containment until 04:00 UTC to allow scoping to continue" | Specificity prevents later reinterpretation |
| Decision maker | Full name and title of the authorizing person | Establishes authority and accountability |
| Information available at the time | What was known when the decision was made — even if later found to be incomplete or wrong | The critical legal protection: decisions are judged against information available then, not information discovered later |
| Alternatives considered | What other options were evaluated and why they were rejected | Shows the decision was thoughtful, not careless |
| Outcome (filled retrospectively) | Did the decision achieve its intended purpose? | Enables lessons learned; documents whether the decision logic was sound |
Every IR involves decisions made under uncertainty with incomplete information. From the outside, months or years later, those decisions can look wrong — because by then, you know things you didn't know then. The legal and regulatory standard for evaluating IR decisions is reasonableness given the information available at the time — not omniscience. A decision to not isolate a host because the blast radius was unknown, documented as such with the specific information available at that moment, is a defensible decision. The same decision, undocumented, looks like reckless inaction. The decision log entry converts uncertainty and risk management into documented professional judgment. Write what you knew, not what you later learned.
Evidence Inventory and Chain of Custody
Forensic evidence is only useful in legal proceedings if chain of custody is maintained. An unbroken chain of custody means you can prove that a piece of evidence was not tampered with between collection and presentation. Breaking the chain doesn't necessarily mean evidence is inadmissible, but it creates grounds for challenge that you do not want in a high-stakes case.
| Evidence inventory field | What to record |
|---|---|
| Evidence ID | Unique identifier: EVD-001, EVD-002 |
| Description | What it is: "Memory image of LAPTOP-JSMITH", "Malware sample extracted from C:\Users\Public\evil.exe" |
| Source | Hostname, IP, share path, user account — where it came from exactly |
| Collection time (UTC) | When it was collected |
| Collected by | Full name of analyst |
| Collection method | Tool used: WinPmem 3.3.rc3, FTK Imager 4.7, manual copy |
| Hash (SHA256) | Computed immediately on collection, before any analysis — this is the integrity anchor |
| Storage location | Exact path on the evidence server or external drive |
| Access log | Every person who accessed the evidence after collection, with timestamp and purpose |
The Three-Audience Final Report
A single incident generates documentation needs for three completely different audiences. The mistake organizations make is writing one technical report and sending it to everyone. Executives cannot use a 40-page technical timeline. Lawyers cannot work with a document full of undefined acronyms. Technical staff cannot do post-incident work from a two-page business summary. Write three separate documents from the same underlying facts.
Final Incident Report Structure
══════════════════════════════════════════════════════════════════════
DOCUMENT 1: Executive Summary (2-3 pages)
──────────────────────────────────────────
Audience: CISO, CEO, Board (may be presented in board meeting)
Contents:
• What happened (2-3 sentences, no jargon)
• Business impact (systems affected, downtime, estimated cost)
• How it was detected (and if detection was delayed, why)
• What was done (high-level response actions)
• Current status (resolved? monitoring? ongoing?)
• What we are doing to prevent recurrence (2-3 bullets)
Forbidden: Technical detail, acronyms without definition, speculation
DOCUMENT 2: Legal and Regulatory Section
─────────────────────────────────────────
Audience: General counsel, outside counsel, compliance, insurance
Contents:
• Data types confirmed or potentially accessed
• Evidence of access vs evidence of exfiltration (clearly separated)
• Regulatory notification analysis (what triggers apply, by when)
• Timeline of key decisions and notifications (with legal relevance)
• Security controls that were in place at time of breach
• Evidence of reasonable security practices
Forbidden: Speculation about attacker motive, legal conclusions
(these belong to counsel, not the IR team)
DOCUMENT 3: Technical Appendix (as many pages as needed)
─────────────────────────────────────────────────────────
Audience: Security operations, IT, future IR teams
Contents:
• Complete technical timeline with evidence references
• Attack chain reconstruction (initial access → persistence →
lateral movement → impact)
• MITRE ATT&CK technique mapping
• All IOCs with context
• Detection gaps (what should have fired and didn't)
• New detection rules recommended
• Hardening recommendations (specific, not generic)
• Tool and process gaps identified during IR
What Not to Write
Everything written during an incident — including Slack messages, emails, and informal notes — is potentially discoverable in litigation. The discipline of writing correctly during high-stress situations requires specific habits.
| What not to write | Why | Write this instead |
|---|---|---|
| "This happened because we've been negligent about patching for years" | Legal conclusion + admission of negligence — discoverable and damaging | "The initial access vector was CVE-2024-XXXX, patched in [date]. We are reviewing patch management procedures." |
| "We really dropped the ball on this one" | Admission that could be used against the organization or individual in litigation | Document the specific gaps and their causes factually, without characterizing fault |
| "I think they probably got our customer database" | Speculation presented as assessment — may be used as admission before investigation is complete | "Investigation into potential data access to the customer database is ongoing. Findings will be documented when complete." |
| "Legal told us not to document this" | Documenting that you were told not to document something is worse than the documentation itself — it suggests intentional destruction of evidence | If legal provides guidance about documentation, follow it. Do not document the instruction itself in the incident record. |
| "[Name] screwed this up by not patching that server" | Blame attribution — creates personnel liability and is rarely accurate (individual blame for systemic failures) | "Server X was unpatched due to [documented process gap]. Remediation: [specific fix]." |
This is not paranoia — it is accurate. Any document created during or about an incident is potentially subject to legal discovery. The practical implication: write facts, not opinions. Write what you observed, not what you infer about fault or negligence. Write what you knew at the time, not what later turned out to be true. If you are in a Slack channel discussing the incident with your team, assume that channel will be read by outside counsel. The question to ask before writing anything is: "If this appears in a courtroom in 18 months, does it reflect well on the organization's response?" Professional, factual, evidence-based documentation passes that test. Speculation, blame, and informal commentary do not.
Q & A
Q: An analyst sends a Slack message during the incident saying "I think this happened because [specific system] was never secured properly." The statement is factually incorrect — the breach had nothing to do with that system. What do you do?
Act on it quickly: (1) Correct the record. Have the analyst or the IR lead post a factual correction in the same Slack thread: "Clarification: our investigation has not identified [specific system] as related to this incident. The initial access vector is [confirmed vector or 'still under investigation']." A correction in the same thread is better than leaving the incorrect statement standing. (2) Brief legal counsel. The incorrect statement exists, and counsel needs to know about it so they can address it if it comes up in discovery. They will not be surprised later — much better. (3) After the incident: reinforce the IR communication protocol. All technical findings should be documented in the case file, not in general communication channels. Slack statements during incidents should be operational coordination, not forensic analysis. (4) Do not delete the message. Intentional deletion of potentially relevant communications during a litigation hold period is spoliation — destruction of evidence — which carries its own legal consequences far worse than the original message. Correct it, document the correction, and tell legal it exists.
Q: The IR was completed by contractors who are no longer engaged. How do you produce a final incident report after the fact when the knowledge is dispersed?
This is a common post-incident problem and it illustrates why documentation during the incident is critical. The recovery approach: (1) Collect all available artifacts — the SIEM search history, EDR queries run during the IR, any evidence the contractors collected, email threads, and any notes they left. (2) Interview the contractors if they are reachable — a 1-hour debrief call with good notes is better than nothing. Many contractors are willing to provide a short post-incident debrief. (3) Reconstruct the timeline from log data rather than memory — SIEM audit logs, EDR query history, and case management system activity logs can reconstruct the sequence of events even without human notes. (4) Write what you can reconstruct with confidence, and clearly document the gaps: "Activities between T+4 and T+8 hours are not fully reconstructable due to absence of contemporaneous notes." A partial but honest report is better than a fabricated complete one. (5) The lessons learned finding from this incident should be: require contractors to document in a company-controlled system from day one, with handoff procedures specified in the contract.
Q: Legal wants a copy of all forensic images for discovery. The images total 4TB across 15 hosts. How do you handle the legal discovery process for forensic evidence?
Forensic evidence discovery has specific protocols that differ from normal document production: (1) Do not hand over forensic images directly. Forensic images contain everything on a disk, including personal employee data, privileged communications, and data unrelated to the incident. Providing raw forensic images to litigation parties may create separate privacy and privilege issues. (2) Work with outside counsel on a targeted production approach — they typically request that a forensic examiner extract and produce specific artifacts (log files, specific directories, malware samples) rather than raw images. (3) Maintain chain of custody throughout. Every transfer of evidence must be documented: when it was transferred, to whom, using what method (encrypted transfer, physical media with tracking), with hash verification at both ends. (4) Images should be produced as verified forensic copies — hash-verified copies, not originals. The originals remain in your custody for your own continued use. (5) Engage e-discovery counsel (a specialty within law — different from general outside counsel) early in the process. They understand the technical nuances of forensic evidence production and will protect you from inadvertently producing privileged or over-broad materials.
Q: The case management system you used for the incident is a SaaS tool, and the vendor announced end-of-life. How do you preserve the incident record?
Export everything now, before the platform is shut down: (1) Export the full case record in all available formats — most case management tools offer JSON, PDF, or CSV export. Export everything: cases, tasks, observables, timeline, attachments. (2) Hash the exported archive immediately on creation. This establishes integrity — you can prove the export was not modified after the platform's shutdown. (3) Store the export in long-term archival storage (not just the IR team's personal drives): a document management system, a compliant cloud storage bucket, or physical media in secure storage, depending on your retention requirements. (4) Document the migration — note in the incident record that the case was preserved in a specific format and location after platform end-of-life, and why. (5) Translate key artifacts into format-independent storage where possible — a timeline that only exists in the SaaS platform's proprietary format is less durable than the same timeline in a Markdown file. Incident records should be stored in formats that can be read without the original tool.
Q: The board wants a post-incident presentation. How much technical detail do you include?
Minimal, framed in business terms. The board's concerns are: business impact (quantified if possible), whether the organization acted responsibly, whether the organization is protected going forward, and whether additional investment is required. They are not concerned with technical attack chains unless you are making a budget argument that requires the context. Structure the board presentation as: (1) What happened — two sentences, plain English. (2) Business impact — systems affected, downtime in hours, data at risk, estimated cost including IR costs and business interruption. (3) How we responded — timeline in business terms ("within 4 hours we had contained the breach; within 72 hours systems were being restored") not technical terms. (4) Why it happened — one root cause, stated without blame, framed as a systemic issue. "Our endpoint detection coverage had a gap on servers in the backup VLAN" not "Joe didn't patch the server." (5) What we're doing about it — three to five specific, actionable items with owners and timelines. This is your opportunity to secure budget for the remediation. (6) Current security posture — one slide showing where things stand now relative to before the incident. The presentation should take 15 minutes. The board does not want more than that, and more than that suggests you're padding to avoid the hard questions.
Q: You are working the IR and you discover evidence that a senior executive was involved in the initial compromise — either through negligence or complicity. How does documentation change?
This is a situation that requires immediate escalation to the general counsel and, depending on your role and jurisdiction, potentially to the board directly — bypassing the executive chain of command if the executive in question is in that chain. Documentation considerations: (1) Ensure the finding is documented factually and specifically in the case file, with full evidence references, before raising it with anyone. You need to be able to support the finding with documented evidence, not a verbal claim. (2) Limit access to this portion of the case file to those with a need to know — this is a personnel matter that, if it becomes a legal matter, will have attorney-client privilege implications. (3) Do not alter, omit, or delay documenting the finding. Evidence that you found something and delayed reporting it is worse than the finding itself. (4) If the organization has a board-level audit committee or risk committee, this type of finding may need to go directly there. (5) Engage outside counsel immediately — independent counsel, not the company's general counsel if there is any possibility the general counsel's loyalty is complicated by the executive's involvement. This situation is beyond the normal scope of IR and requires legal guidance specific to your jurisdiction and circumstances.