| name | ir-report-builder |
| description | Analyze security logs and incident data to reconstruct an attack timeline, build an incident response plan following NIST SP 800-61 / SANS PICERL, and produce a detailed technical report plus a concise executive summary. Use when the user mentions 'IR report,' 'incident response report,' 'incident analysis,' 'log analysis for incident,' 'attack timeline,' 'root cause analysis,' 'executive summary of incident,' 'post-incident report,' 'IR plan,' 'containment plan,' or has raw logs / alerts and needs them turned into an investigation and report. |
| allowed-tools | Read, Grep, Glob, Bash, Write |
IR Report Builder
Turn raw logs, alerts, and incident notes into a defensible investigation: a reconstructed timeline, a response plan mapped to a recognized methodology, and two audiences' worth of reporting — deep technical detail for responders and a tight executive summary for leadership.
Defensive use. This skill is for authorized incident response on systems you own or are engaged to defend. Preserve evidence integrity; never alter source logs — work on copies.
Activation triggers
Use this skill when the user:
- Provides logs (firewall, EDR, auth, web, proxy, DNS, cloud audit) and asks "what happened."
- Asks for an incident report, post-incident review, or executive summary.
- Needs an IR plan, containment/eradication/recovery steps, or a lessons-learned writeup.
- Wants a timeline or root-cause analysis from scattered data.
Methodology
Follow NIST SP 800-61r2 phases, cross-referenced with SANS PICERL:
Preparation → Detection & Analysis → Containment → Eradication → Recovery → Post-Incident.
Map adversary behavior to MITRE ATT&CK technique IDs wherever the evidence supports it.
Workflow
-
Intake & scope. Ask for (or infer from provided data): incident type, affected assets, detection time, data sources available, and business criticality. Confirm the reporting window and time zone. Never assume authorization — confirm it.
-
Normalize the evidence. Identify each log source and its format. Use scripts/log_timeline.py to parse common formats (RFC3164/RFC5424 syslog, JSON lines, CSV, CEF, Apache/Nginx access, FortiGate & appliance key-value/logfmt, and AWS CloudTrail JSON) into a single normalized, timezone-aligned event list. Keep a hash of every source file for chain of custody.
-
Reconstruct the timeline. Order events chronologically. Flag the initial access, first malicious action, lateral movement, privilege escalation, impact, and last observed activity. Note gaps in logging as gaps, not as absence of activity.
-
Analyze.
- Extract IOCs (IPs, domains, hashes, user accounts, file paths) into
resources/ioc-extract.md shape.
- Map each observed step to a MITRE ATT&CK technique.
- Determine root cause and the exploited weakness.
- Assess blast radius: what was accessed, exfiltrated, or altered.
-
Build the IR plan. Produce concrete Containment / Eradication / Recovery actions with owners and priority. Distinguish already done vs recommended next.
-
Write two deliverables.
- Technical report → fill
templates/ir-report-template.md. Dense, evidence-cited, reproducible.
- Executive summary → fill
templates/exec-summary-template.md. One page, business-impact-first, no jargon. Answer: what happened, what's the impact, is it contained, what do we need from leadership.
-
Quality gate. Every claim in the report must cite an evidence line (timestamp + source). Mark anything inferred as assessed with [high/medium/low] confidence. List logging/visibility gaps explicitly.
Output format
- Technical report: Markdown, sectioned per the template, with an appendix of raw evidence excerpts and IOC tables.
- Executive summary: single page, leads with impact and status. Use a severity label (Critical/High/Medium/Low) and a one-line "bottom line up front."
Handoffs
- IOCs and observed techniques → feed into
siem-detection-engineer to build detections that would have caught this earlier.
- Containment actions that should be automated → feed into
soar-playbook-builder.