Extended attribute syscall or setfattr/getfattr execution on Linux

Siemphony@siemphonylowT1564.014unverifiedPlaybook · Defence evasion

Matches AN1135's Linux chain at the two points a single auditd record can carry it: a SYSCALL record naming one of the six xattr syscalls (setxattr, lsetxattr, fsetxattr and their get counterparts), or an EXECVE record whose argv0 is setfattr or getfattr. Both are real, single-event observables -- writing or reading an extended attribute always goes through one of the six syscalls, and the CLI tools always execve into a process named for themselves. AN1135's full three-part chain -- a namespace-scoped setfattr, a size or entropy anomaly in the value written, and a later process reading that value back -- collapses to the first two steps here: a raw auditd SYSCALL record carries no decoded attribute name or value, only numeric syscall arguments (register contents, not parsed strings), so MITRE's XattrNamespaces and PayloadSizeThreshold knobs cannot be applied to it, and "a later process reads the value back" is a cross-event correlation within MITRE's CorrelationWindow that lib/sigma has no join or timeframe construct to express. The EXECVE selection is similarly left unscoped to a namespace: setfattr's `-n user.foo` argument can land in any of several argv positions depending on what else is on the command line, and pinning one position would silently miss the others. AN1136, the macOS leg, is not attempted: its logsource, {product: macos, service: unifiedlog}, has no standardised Sigma field vocabulary anywhere in this corpus, the same blocker already recorded against it for T1036.004, T1552.003, T1564.001 and others. Prerequisite: raw auditd audits neither xattr syscalls nor process execution by default. This rule returns zero rows until both an xattr watch (`-a always,exit -F arch=b64 -S setxattr,lsetxattr,fsetxattr,getxattr,lgetxattr,fgetxattr -k xattr`) and an execve watch (`-a always,exit -F arch=b64 -S execve -k exec`) are loaded -- and zero rows here reads as quiet when it actually means blind. Second prerequisite specific to `selection_xattr_syscall`'s `syscall` field: raw, uninterpreted auditd output records `syscall=` as the architecture-dependent kernel syscall number, not the name string this selection matches against (`setxattr` is 188 on x86_64, for example, and the numbering differs on arm64). Name resolution is not universal -- Elastic's Auditbeat resolves it by default, but a plain-text `audit.log` tail, `ausearch` without `-i`, and Wazuh's default auditd field extraction all pass the number through unresolved, so on those collectors this half of the rule never fires until an interpreting layer (`ausearch -i`, `laurel`, or an equivalent enrichment pipeline) sits in front of it. `selection_xattr_cmd`'s `a0` is unaffected by this -- EXECVE always records argv as text, never as a number, regardless of collector. UNVERIFIED AS AUTHORED — derived from MITRE ATT&CK DET0406, not hand-written and not tested by its author. Any lab result is on this rule's own page.Full description

The 3 SIEM queries below are previews produced by Siemphony’s own translator, not pySigma, and none has been executed against a real backend — review before you deploy.

detection.yml

1title: Extended attribute syscall or setfattr/getfattr execution on Linux2id: 8dcc029a-0a2e-4ae4-9c8e-195081030fc23status: experimental4description: |5  Matches AN1135's Linux chain at the two points a single auditd record can6  carry it: a SYSCALL record naming one of the six xattr syscalls (setxattr,7  lsetxattr, fsetxattr and their get counterparts), or an EXECVE record whose8  argv0 is setfattr or getfattr. Both are real, single-event observables --9  writing or reading an extended attribute always goes through one of the six10  syscalls, and the CLI tools always execve into a process named for11  themselves. AN1135's full three-part chain -- a namespace-scoped setfattr,12  a size or entropy anomaly in the value written, and a later process reading13  that value back -- collapses to the first two steps here: a raw auditd14  SYSCALL record carries no decoded attribute name or value, only numeric15  syscall arguments (register contents, not parsed strings), so MITRE's16  XattrNamespaces and PayloadSizeThreshold knobs cannot be applied to it, and17  "a later process reads the value back" is a cross-event correlation within18  MITRE's CorrelationWindow that lib/sigma has no join or timeframe construct19  to express. The EXECVE selection is similarly left unscoped to a namespace:20  setfattr's `-n user.foo` argument can land in any of several argv21  positions depending on what else is on the command line, and pinning one22  position would silently miss the others. AN1136, the macOS leg, is not23  attempted: its logsource, {product: macos, service: unifiedlog}, has no24  standardised Sigma field vocabulary anywhere in this corpus, the same25  blocker already recorded against it for T1036.004, T1552.003, T1564.00126  and others.27  Prerequisite: raw auditd audits neither xattr syscalls nor process28  execution by default. This rule returns zero rows until both an xattr29  watch (`-a always,exit -F arch=b64 -S30  setxattr,lsetxattr,fsetxattr,getxattr,lgetxattr,fgetxattr -k xattr`) and an31  execve watch (`-a always,exit -F arch=b64 -S execve -k exec`) are loaded --32  and zero rows here reads as quiet when it actually means blind.33  Second prerequisite specific to `selection_xattr_syscall`'s `syscall` field:34  raw, uninterpreted auditd output records `syscall=` as the35  architecture-dependent kernel syscall number, not the name string this36  selection matches against (`setxattr` is 188 on x86_64, for example, and37  the numbering differs on arm64). Name resolution is not universal --38  Elastic's Auditbeat resolves it by default, but a plain-text `audit.log`39  tail, `ausearch` without `-i`, and Wazuh's default auditd field extraction40  all pass the number through unresolved, so on those collectors this half of41  the rule never fires until an interpreting layer (`ausearch -i`, `laurel`,42  or an equivalent enrichment pipeline) sits in front of it.43  `selection_xattr_cmd`'s `a0` is unaffected by this -- EXECVE always records44  argv as text, never as a number, regardless of collector.45  UNVERIFIED AS AUTHORED — derived from MITRE ATT&CK DET0406, not hand-written and not tested by its author. Any lab result is on this rule's own page.46references:47  - https://attack.mitre.org/techniques/T1564/01448  - https://attack.mitre.org/detectionstrategies/DET040649author: Siemphony pipeline (machine-derived from MITRE ATT&CK v19.2)50date: 2026-08-1851tags:52  - attack.defense-evasion53  - attack.t1564.01454logsource:55  product: linux56  service: auditd57detection:58  selection_xattr_syscall:59    type: 'SYSCALL'60    syscall:61      - 'setxattr'62      - 'lsetxattr'63      - 'fsetxattr'64      - 'getxattr'65      - 'lgetxattr'66      - 'fgetxattr'67  selection_xattr_cmd:68    type: 'EXECVE'69    a0|endswith:70      - 'setfattr'71      - 'getfattr'72  condition: 1 of selection*73falsepositives:74  - "SELinux relabeling and policy tooling -- restorecon, setfiles and semanage -- calling get/setxattr against security.selinux on every file touched during a full relabel, an fixfiles run, or routine file creation on any enforcing SELinux host; this is the highest-volume match by a wide margin and is why this rule is low rather than medium."75  - "Backup, archive and sync tooling that preserves extended attributes -- tar with --xattrs, rsync -X, bsdtar and cp --preserve=xattr -- each of which calls get/setxattr once per file it copies or archives, producing one burst of matches per scheduled job."76  - "Administrators and provisioning scripts running getfattr or setfattr directly to inspect or set POSIX ACLs (system.posix_acl_access) or an SELinux context as routine filesystem administration, which is the exact command this rule's selection_xattr_cmd is built to match."77  - "Container and package-management tooling -- dockerd/containerd extracting image layers, and rpm/dpkg installing packages that carry Linux file capabilities -- calling setxattr on security.capability for every capability-bearing binary during a pull, build or install."78level: low

Splunk · SPL

Run this as a search.

index=* ((type="SYSCALL" AND (syscall="setxattr" OR syscall="lsetxattr" OR syscall="fsetxattr" OR syscall="getxattr" OR syscall="lgetxattr" OR syscall="fgetxattr")) OR (type="EXECVE" AND (a0="*setfattr" OR a0="*getfattr")))

Elastic · ES|QL

Run this as a search.

FROM logs-*| WHERE ((TO_LOWER(type) == "syscall" AND (TO_LOWER(syscall) == "setxattr" OR TO_LOWER(syscall) == "lsetxattr" OR TO_LOWER(syscall) == "fsetxattr" OR TO_LOWER(syscall) == "getxattr" OR TO_LOWER(syscall) == "lgetxattr" OR TO_LOWER(syscall) == "fgetxattr")) OR (TO_LOWER(type) == "execve" AND (TO_LOWER(a0) LIKE "*setfattr" OR TO_LOWER(a0) LIKE "*getfattr")))

Wazuh · XML rule

Deploy to your manager — this is a rule, not a search.

<group name="sigma,linux,">  <!-- Rule ids must be unique on your manager; 100000+ is the user range. -->  <!-- 2 rules: the Sigma condition ORs across different fields,       which one rule cannot express. Any one matching is a hit. -->  <rule id="100000" level="5">    <!-- Set <if_sid> to the decoder/base rule for linux so this only evaluates relevant events. -->    <field name="type" type="pcre2">(?i)^SYSCALL$</field>    <field name="syscall" type="pcre2">(?i)(^setxattr$|^lsetxattr$|^fsetxattr$|^getxattr$|^lgetxattr$|^fgetxattr$)</field>    <description>Extended attribute syscall or setfattr/getfattr execution on Linux (1/2)</description>    <mitre>      <id>T1564.014</id>    </mitre>  </rule>   <rule id="100001" level="5">    <!-- Set <if_sid> to the decoder/base rule for linux so this only evaluates relevant events. -->    <field name="type" type="pcre2">(?i)^EXECVE$</field>    <field name="a0" type="pcre2">(?i)(setfattr$|getfattr$)</field>    <description>Extended attribute syscall or setfattr/getfattr execution on Linux (2/2)</description>    <mitre>      <id>T1564.014</id>    </mitre>  </rule></group>

Sentinel · KQL

Run this as a search.

Sentinel cannot express this construct. No Sentinel table is modelled for this logsource ({"product":"linux","service":"auditd"}), so there is nothing honest to query. Pick a logsource this translator maps, or write the KQL by hand. The Sigma source is on the first tab, unchanged.

At a glance

Logsource: auditd / linux
Level: low
Category: hide_artifacts
Sigma status: experimental
Backends: 3 of 4 rendered
Updated: 30 Aug 2026
Rule ID: 8dcc029a-0a2e-4ae4-9c8e-195081030fc2

ATT&CKT1564.014

T1564.014Extended Attributes

ATT&CK® is a registered trademark of The MITRE Corporation. These mappings are the rule author’s, and are neither reviewed nor endorsed by MITRE.

Evidence

Verification1 run

Not verified

Siemphony benign baseline
completed
Siemphony IR · 3 Sept 2026
No events matched, so this run proves nothing either way.

Noise check0 of 5 fired

Fires on none of the 5 ordinary behaviours the baseline holds for this log type. That is not a precision guarantee: it means only not caught by these 5 behaviours, and the next ordinary behaviour nobody wrote down may still match.

Run live against 65 hand-written events of ordinary corporate activity — backup agents, software distribution, patching, helpdesk scripts, EDR. It says nothing about how often the rule would fire in production, which is the number that actually exhausts an analyst.

Test it yourselfin your browser

Runs here in your browser against public attack recordings. A preview — not the verification badge.

Recordings of T1564.014, evaluated against their — events.

Using it

Tuning & false positivesexperimental · 4 declared

experimental (current stage)
test
stable

Written and reasoned about, but not yet run against real telemetry. Expect to tune it before trusting it.

Tuning tips

Neither selection binds to the audit key= set by the watches above, so once loaded, both selections also match SYSCALL/EXECVE records from any unrelated auditing on the host — a broad -S all policy, or a STIG execve watch. Adding key: 'xattr' and key: 'exec' to the two selections scopes hits to only the watches this rule's own prerequisite installs. MITRE's XattrNamespaces and PayloadSizeThreshold knobs cannot be applied here at all — a raw SYSCALL event has no decoded namespace or value to filter on — and CorrelationWindow would require a join across the xattr write and the later reading process that lib/sigma has no construct to express.

Declared by the author

SELinux relabeling and policy tooling -- restorecon, setfiles and semanage -- calling get/setxattr against security.selinux on every file touched during a full relabel, an fixfiles run, or routine file creation on any enforcing SELinux host; this is the highest-volume match by a wide margin and is why this rule is low rather than medium.
Backup, archive and sync tooling that preserves extended attributes -- tar with --xattrs, rsync -X, bsdtar and cp --preserve=xattr -- each of which calls get/setxattr once per file it copies or archives, producing one burst of matches per scheduled job.
Administrators and provisioning scripts running getfattr or setfattr directly to inspect or set POSIX ACLs (system.posix_acl_access) or an SELinux context as routine filesystem administration, which is the exact command this rule's selection_xattr_cmd is built to match.
Container and package-management tooling -- dockerd/containerd extracting image layers, and rpm/dpkg installing packages that carry Linux file capabilities -- calling setxattr on security.capability for every capability-bearing binary during a pull, build or install.

Lessonhide artifacts

hide artifactsAI draft — unreviewed

01 · What this rule detects

This rule flags two single-event observables from AN1135's Linux xattr-abuse chain: a raw auditd SYSCALL record naming one of six extended-attribute syscalls, or an EXECVE record whose argv0 is setfattr or getfattr. It does not detect payload hiding itself — MITRE's full chain needs a namespace-scoped write, a size or entropy anomaly in the value, and a later process reading it back, and a single auditd event carries none of that context. What ships is a sensor for "something touched an extended attribute or ran the CLI tool for one," not a hiding detector.

02 · Telemetry you need

Raw auditd audits neither xattr syscalls nor process execution by default. Two watches must be loaded before this rule can produce anything: an xattr watch (-a always,exit -F arch=b64 -S setxattr,lsetxattr,fsetxattr,getxattr,lgetxattr,fgetxattr -k xattr) and an execve watch (-a always,exit -F arch=b64 -S execve -k exec). Without both, the rule returns zero rows, and that silence reads as "quiet" when it actually means "blind." A second, narrower prerequisite applies only to selection_xattr_syscall: raw auditd records syscall= as a numeric kernel syscall number, not the name string this selection matches against, and name resolution is not universal — Auditbeat resolves it by default, but a plain audit.log tail, ausearch without -i, and Wazuh's default field extraction all pass the number through unresolved. On those collectors this leg never fires until something like ausearch -i or laurel sits in front of it. selection_xattr_cmd's a0 field is unaffected, since EXECVE always records argv as text.

03 · How the logic works

Two independent selections, ORed by condition: 1 of selection*. selection_xattr_syscall matches type: SYSCALL records where syscall is one of the six xattr get/set calls. selection_xattr_cmd matches type: EXECVE records where a0 ends with setfattr or getfattr. Neither selection is scoped to a namespace: MITRE's AN1135 chain calls out user. and trusted. specifically, but a raw SYSCALL record carries only numeric register arguments, not a decoded attribute name, so there's nothing in that event to filter on. The command-line leg could in principle look for -n user. or -n trusted. in argv, but that argument can land in any position depending on what else is on the command line, so pinning one position would silently miss the others — left unscoped rather than made falsely precise.

04 · Common false positives

The description names the syscall leg as loudest by a wide margin: on any enforcing SELinux host, restorecon, setfiles and semanage call get/setxattr against security.selinux on every file touched during a relabel. Backup and archive tooling that preserves xattrs — tar --xattrs, rsync -X, cp --preserve=xattr — fires once per file copied. Direct administrative use of getfattr/setfattr for ACL or SELinux-context work matches the command leg by design, since that's the exact behavior it's built to catch. Container runtimes and package managers (dockerd, containerd, rpm, dpkg) call setxattr on security.capability for every capability-bearing binary during a pull or install.

05 · Detection maturity

tier_1: This rule has never been run against a log line and sits at the lowest DeBMM tier by default. It parses and renders, but "renders" and "detects" are different claims — this one only has evidence for the first until a human runs it against real auditd output and validates precision against the false positives named above.

Share this rule

/u/siemphony/8dcc029a-0a2e-4ae4-9c8e-195081030fc2

More from this repertoire

All compositions

Verdicts · reactions · comments

Community

Verdicts from engineers who actually deployed it, and the conversation around it.

Not yet reported on

Nobody has run this in a real environment and said what happened.

Verdicts from engineers who deployed it

0 cast

No verdicts reported yet. The first one is worth more than the tenth — say what you ran it against.

No comments yet. Someone who deploys this will have something to say about it.

Extended attribute syscall or setfattr/getfattr execution on Linux

The detection

detection.yml

Splunk · SPL

Elastic · ES|QL

Wazuh · XML rule

Sentinel · KQL

Community

Verdicts from engineers who deployed it