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.
====== Sandbox Evasion Indicators ======

===== Definition =====
Sandbox Evasion Indicators are an anti-analysis or malware behavior analysts examine to understand evasive execution, hidden runtime state, or suspicious code activity.
In a forensic report, sandbox evasion indicators 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 =====
Malware can avoid sandboxes, hide in memory, and make tools earn their keep. The trick is to document behavior without turning every evasive feature into a motive speech.
Good analysis keeps the artifact in its lane: useful, limited, and not allowed to perform intent for the room. Hidden code in memory is serious. It still needs process, host, and timeline context before the report starts naming ghosts.

===== Technical Description =====
The review compares static indicators, sandbox behavior, memory evidence, process lineage, network traffic, and endpoint telemetry.
If the tool normalized, decoded, or reconstructed the record, that transformation belongs in the notes.

===== Forensic Relevance =====
  * **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 sandbox evasion indicators looks interesting but not conclusive.
  * **Timeline reconstruction:** Sandbox Evasion Indicators can help place activity in sequence when the time source and collection conditions are understood.
  * **Recovery analysis:** It can explain why data was recovered, missed, corrupted, or only partially reconstructed.

===== Evidence Sources =====
^ Evidence source ^ What it may show ^ Reliability limits ^ What it cannot prove alone ^
| Executable or script sample | May show imports, strings, capabilities, packers, or obfuscation. | Static features can be misleading or intentionally confusing. | Do not prove deployment context. |
| Sandbox output | May show runtime behavior, files, registry changes, and network activity. | Sandbox awareness and environment mismatch can distort behavior. | Does not prove what happened on the endpoint. |
| Memory and process evidence | May show injected code, command lines, sockets, and modules. | Collection timing matters. | Does not identify the actor. |
| Detection rules and alerts | May support triage and hunt leads. | False positives and rule scope must be documented. | Do not replace artifact analysis. |

===== Interpretation Limits =====
The most common error is treating sandbox evasion indicators 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.
Anti-debugging checks are behavior, not biography.

===== Common Misinterpretations =====
  * Ignoring benign explanations because the suspicious explanation is easier to write.
  * Treating sandbox evasion indicators 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.

===== Example Scenario =====
An examiner finds executable or script sample related to sandbox evasion indicators during a triage review.
The record may be meaningful, but it has to be compared with sandbox output before the report gets confident.
The report should state the observed record, the collection limits, and the alternative explanations that were tested.

===== Analysis Workflow =====
  - Define the question before opening another parser: what should Sandbox Evasion Indicators help answer?
  - Preserve the source evidence and document how the executable or script sample 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 sandbox evasion indicators 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 executable or script sample is consistent with activity related to sandbox evasion indicators, 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 | Executable or script sample exists, but collection scope, time source, or surrounding context is limited. | State the observation and keep interpretation narrow. |
| Moderate | Executable or script sample aligns with Sandbox output, 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:anti_forensics:detection:anti_debugging_indicators|Anti-Debugging Indicators]] - Often appears nearby in reviews of the same system or behavior.
  * [[wiki:anti_forensics:detection:process_injection_traces|Process Injection Traces]] - Helps keep the finding grounded in a wider evidence pattern.
  * [[wiki:anti_forensics:detection:self_deleting_malware|Self-Deleting Malware]] - Useful when checking whether another artifact supports the same interpretation.
  * [[wiki:anti_forensics:detection:alternate_data_stream_abuse|Alternate Data Stream Abuse]] - Helps keep the finding grounded in a wider evidence pattern.
  * [[wiki:anti_forensics:detection:anti_forensic_file_wiping|Anti-Forensic File Wiping]] - Useful when checking whether another artifact supports the same interpretation.
  * [[wiki:anti_forensics:detection:anti_forensics|Anti-Forensics]] - Often appears nearby in reviews of the same system or behavior.

===== Tools =====
  * [[https://github.com/volatilityfoundation/volatility3|Volatility 3]] - Memory analysis and plugin-driven investigation.
  * [[https://remnux.org/|REMnux]] - Linux malware analysis environment.
  * [[https://github.com/mandiant/flare-vm|FLARE-VM]] - Windows malware analysis toolkit.
  * [[https://ghidra-sre.org/|Ghidra]] - Reverse engineering suite.
  * [[https://virustotal.github.io/yara/|YARA]] - Pattern matching for malware triage.
  * [[https://github.com/mandiant/capa|capa]] - Capability detection for executable analysis.
  * [[https://github.com/Velocidex/WinPmem|WinPmem]] - Windows memory acquisition.
  * [[https://github.com/504ensicsLabs/LiME|LiME]] - Linux memory acquisition.
  * [[https://github.com/microsoft/avml|AVML]] - Linux memory acquisition for cloud and endpoint response.
  * [[https://github.com/ufrisk/MemProcFS|MemProcFS]] - Memory analysis through a virtual file-system interface.
  * [[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://volatility3.readthedocs.io/|Volatility 3 documentation]] - Memory analysis documentation.
  * [[https://yara.readthedocs.io/|YARA documentation]] - YARA rule documentation.
  * [[https://www.cisa.gov/resources-tools|CISA resources and tools]] - CISA cybersecurity resources and tools.
  * [[https://csrc.nist.gov/publications/detail/sp/800-86/final|NIST SP 800-86]] - Forensic and incident response handling guidance.
  * [[https://www.forensicfocus.com/articles/countering-anti-forensic-efforts-part-1/|Forensic Focus: Countering Anti-Forensic Efforts]] - Defensive anti-forensics discussion.

===== See Also =====
  * [[wiki:anti_forensics:detection:start|Detection]]
  * [[wiki:anti_forensics:detection:anti_debugging_indicators|Anti-Debugging Indicators]]
  * [[wiki:anti_forensics:detection:process_injection_traces|Process Injection Traces]]
  * [[wiki:anti_forensics:detection:self_deleting_malware|Self-Deleting Malware]]
  * [[wiki:anti_forensics:detection:alternate_data_stream_abuse|Alternate Data Stream Abuse]]
  * [[wiki:anti_forensics:detection:anti_forensic_file_wiping|Anti-Forensic File Wiping]]
  * [[wiki:anti_forensics:detection:anti_forensics|Anti-Forensics]]
  * [[wiki:anti_forensics:detection:artifact_deletion|Artifact Deletion]]
  * [[wiki:anti_forensics:detection:artifact_gaps|Artifact Gaps]]
  * [[wiki:anti_forensics:detection:benign_explanations_in_anti_forensics|Benign Explanations in Anti-Forensics]]
  * [[wiki:anti_forensics:detection:browser_anti_forensics|Browser Anti-Forensics]]
  * [[wiki:tools:start|Forensic Tools]]
  * [[wiki:resources:start|Project Resources]]

===== Reader Takeaway =====
Sandbox Evasion Indicators are useful when it helps explain what the evidence can support and where the limits begin. Treat sandbox evasion indicators as one part of a corroborated record, not a shortcut to intent. Evasive malware can dodge a sandbox without confessing who brought it to the endpoint.

===== 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.
Do not use this material to destroy, conceal, or tamper with evidence.