Anti-Forensics Field Manual

This page is read only. You can view the source, but not change it. Ask your administrator if you think this is wrong.
====== SRUM ======

===== Definition =====
SRUM is a Windows artifact or artifact family used to interpret execution, user activity, application behavior, device use, or operating-system state.
In a forensic report, srum should be treated as a source of evidence and uncertainty, not as a shortcut to intent. The useful question is what the record supports, what it does not support, and what another source says when asked the same question.

===== Background =====
Windows leaves a large evidence trail, but the trail was not designed to be polite. Some records are loud, some are stale, and some look guilty because the operating system has terrible social skills.
Good analysis keeps the artifact in its lane: useful, limited, and not allowed to perform intent for the room. The artifact may know a path and a time. It does not know motive.

===== Technical Description =====
Analysis usually compares artifact records with user SIDs, file paths, timestamps, execution context, registry state, event logs, and collection notes.
A careful review starts with the source format and collection method before it argues about meaning.

===== Forensic Relevance =====
  * **Recovery analysis:** It can explain why data was recovered, missed, corrupted, or only partially reconstructed.
  * **User activity review:** It may support account or profile context, but only when independent artifacts point the same direction.
  * **Contradiction testing:** It is useful for finding places where logs, metadata, storage behavior, or accounts disagree.
  * **Reporting decisions:** It helps decide how narrow the finding should be when srum looks interesting but not conclusive.
  * **Timeline reconstruction:** SRUM can help place activity in sequence when the time source and collection conditions are understood.

===== Evidence Sources =====
^ Evidence source ^ What it may show ^ Reliability limits ^ What it cannot prove alone ^
| SRUM database | May show application resource usage, network usage, and user SID context. | Retention and service behavior affect what remains. | Does not prove why the application ran. |
| Network telemetry | May corroborate destination and timing. | May be shared by processes or users. | Does not prove content. |
| Event logs | May support account and system context. | Logging coverage varies. | Does not replace SRUM interpretation. |
| User profile artifacts | May tie activity to a profile. | Shared systems complicate attribution. | Does not prove the person at the keyboard. |

===== Interpretation Limits =====
The most common error is treating srum as proof of motive. It may support a technical event, a sequence, or a contradiction, but motive needs stronger ground.
Another bad leap is treating absence as intent. Missing data can come from retention, configuration, collection scope, sync behavior, overwriting, media behavior, or ordinary use.
A tool result should be described as a parsed or recovered record, not as the tool's opinion about what happened. Tools surface evidence; they do not understand it.
Windows artifacts keep receipts, but some receipts are faded, duplicated, or from a store nobody visited.

===== Common Misinterpretations =====
  * Letting one Windows artifact carry the whole timeline.
  * Treating path and time as motive.
  * Treating srum as proof that a specific user acted, when the artifact only supports system or account context.
  * Treating a timestamp as exact truth without checking timezone, clock drift, and source semantics.
  * Ignoring benign explanations because the suspicious explanation is easier to write.

===== Example Scenario =====
During follow-up analysis, srum database becomes the strongest visible record tied to srum.
The artifact may explain part of the sequence, but ordinary system behavior still needs to be ruled in or out.
The analyst preserves the source, exports the relevant records, checks network telemetry, and documents what remains unresolved.

===== Analysis Workflow =====
  - Define the question before opening another parser: what should SRUM help answer?
  - Preserve the source evidence and document how the srum database was collected.
  - Record tool versions, input paths, output paths, time settings, and errors.
  - Identify observed facts before writing any interpretation.
  - Normalize time sources and document timezone, clock drift, and collection-time effects.
  - Compare at least two independent artifact families before raising confidence.
  - Consider benign explanations, automated behavior, retention, sync, and storage-device behavior.
  - Write conclusions proportionally: observed fact first, inference second, uncertainty always visible.

===== Reporting Guidance =====
Reporting on srum should be precise enough that another analyst can retrace the claim without inheriting the original examiner's confidence.
Avoid wording that converts possibility into intent. The report should not say a user deliberately deleted, hid, wiped, or tampered with evidence unless the evidence actually supports that conclusion.
Report-ready wording:
  * The available srum database is consistent with activity related to srum, but it does not independently establish motive or user intent.
  * Recovery was limited under the examined conditions; additional corroboration would be required before concluding deliberate destruction.
  * The finding should be read with the collection scope, time-source limits, and alternate explanations described in this report.

===== Confidence and Reliability =====
^ Confidence level ^ What it looks like for this topic ^ How to report it ^
| Low | SRUM database exists, but collection scope, time source, or surrounding context is limited. | State the observation and keep interpretation narrow. |
| Moderate | SRUM database aligns with Network telemetry, but attribution or intent remains unresolved. | Say the artifacts support the finding, not that they prove it. |
| High | Multiple independent sources agree on sequence, source system, account context, and collection conditions. | Use stronger language, but still separate observed facts from inference. |

===== Related Concepts =====
  * [[wiki:digital_forensics:windows:amcache|Amcache]] - Compare this concept when testing sequence and corroboration.
  * [[wiki:digital_forensics:windows:appcompatcache|AppCompatCache]] - Useful when checking whether another artifact supports the same interpretation.
  * [[wiki:digital_forensics:windows:bam_and_dam|BAM and DAM]] - Compare this concept when testing sequence and corroboration.
  * [[wiki:digital_forensics:windows:bits_jobs|BITS Jobs]] - Useful when checking whether another artifact supports the same interpretation.
  * [[wiki:digital_forensics:windows:cloud_clipboard_artifacts|Cloud Clipboard Artifacts]] - Often appears nearby in reviews of the same system or behavior.
  * [[wiki:digital_forensics:windows:iconcache|IconCache]] - Useful when checking whether another artifact supports the same interpretation.

===== Tools =====
  * [[https://ericzimmerman.github.io/#!index.md|Eric Zimmerman's Tools]] - Windows artifact parsers including MFTECmd, PECmd, EvtxECmd, and others.
  * [[https://www.sans.org/tools/kape|KAPE]] - Targeted artifact collection and processing.
  * [[https://learn.microsoft.com/en-us/sysinternals/|Sysinternals Suite]] - Microsoft Windows internals and triage tools.
  * [[https://learn.microsoft.com/en-us/sysinternals/downloads/sysmon|Sysmon]] - High-fidelity Windows endpoint telemetry.
  * [[https://github.com/keydet89/RegRipper3.0|RegRipper]] - Registry parsing framework and plugins.
  * [[https://eventlogxp.com/|Event Log Explorer]] - Windows event log viewing and filtering.
  * [[https://learn.microsoft.com/en-us/sysinternals/downloads/autoruns|Autoruns]] - Persistence and autorun location review.
  * [[https://learn.microsoft.com/en-us/sysinternals/downloads/procmon|Process Monitor]] - Process, registry, and file activity observation.
  * [[https://github.com/Yamato-Security/hayabusa|Hayabusa]] - Windows event log hunting and timeline support.
  * [[https://github.com/WithSecureLabs/chainsaw|Chainsaw]] - Windows event log search and detection.
  * [[wiki:tools:start|Forensic Tools]] - Full tool directory for the wiki.

===== Limitations of Tools =====
Tools parse, surface, and organize evidence. They do not create conclusions.
Parser output can be affected by version differences, unsupported formats, corrupted records, timezone handling, partial collection, and storage behavior outside the tool's view.
When a tool produces a strong-looking result, validate it against another tool or source where the stakes justify it. A parser can recover a fragment; it cannot tell you whether the fragment deserves a paragraph in the report.

===== References and Further Reading =====
  * [[https://learn.microsoft.com/en-us/sysinternals/|Microsoft Sysinternals]] - Windows internals and triage tools.
  * [[https://learn.microsoft.com/en-us/windows/win32/eventlog/event-logging|Microsoft Windows Event Log]] - Event logging reference.
  * [[https://www.forensicfocus.com/articles/windows-forensics-and-security/|Forensic Focus: Windows Forensics and Security]] - Windows artifact overview.
  * [[https://csrc.nist.gov/publications/detail/sp/800-86/final|NIST SP 800-86]] - Guide to integrating forensic techniques into incident response.
  * [[https://www.nist.gov/itl/ssd/software-quality-group/computer-forensics-tool-testing-program-cftt|NIST CFTT]] - Computer Forensics Tool Testing program.

===== See Also =====
  * [[wiki:digital_forensics:windows:start|Windows]]
  * [[wiki:digital_forensics:windows:amcache|Amcache]]
  * [[wiki:digital_forensics:windows:appcompatcache|AppCompatCache]]
  * [[wiki:digital_forensics:windows:bam_and_dam|BAM and DAM]]
  * [[wiki:digital_forensics:windows:bits_jobs|BITS Jobs]]
  * [[wiki:digital_forensics:windows:cloud_clipboard_artifacts|Cloud Clipboard Artifacts]]
  * [[wiki:digital_forensics:windows:iconcache|IconCache]]
  * [[wiki:digital_forensics:windows:jump_lists|Jump Lists]]
  * [[wiki:digital_forensics:windows:lnk_files|LNK Files]]
  * [[wiki:digital_forensics:windows:mounteddevices|MountedDevices]]
  * [[wiki:digital_forensics:windows:ntfs_change_journal|NTFS Change Journal]]
  * [[wiki:tools:start|Forensic Tools]]
  * [[wiki:resources:start|Project Resources]]

===== Reader Takeaway =====
SRUM is useful when it helps explain what the evidence can support and where the limits begin. Treat srum as one part of a corroborated record, not a shortcut to intent. A Windows trace can look guilty and still prove less than the report wants it to prove.

===== Use Notes =====
This article is for defensive education and technical reference. It should not be treated as legal, forensic, investigative, compliance, or operational advice without qualified professional judgment.