Read-Only Allow-Listed Utility Invoked With a File-Write or State-Changing Option in a Command Argument

Detects an agent or MCP tool call whose command-string argument starts with a utility that command allow-lists commonly classify as read-only (sort, uniq, diff, date, hostname, file) and gives it the option or operand that makes the utility itself write a file, run a helper program, or change system state. No shell metacharacter is involved: the allowed binary opens the destination on its own, so filters that only look for ; | && $( ) never see it. This is the shape reported as CVE-2026-105789 / GHSA-85w2-wggf-rw49 against Microsoft UFO before 3.0.9. The execute_command tool of the Linux MCP server (ufo/client/mcp/http_servers/linux_mcp_server.py) accepted a free-form command string and allowed sort and uniq as read-only commands, so 'sort -o <file>' or the optional second uniq operand created or overwrote any file writable by the server process (agent configuration, shell startup files). The fix (commit 4f79362) removed sort and added per-command argument policies; its regression tests (tests/test_linux_mcp_command_policy.py) list the rejected forms used here. The command string is produced by the model, so a prompt injection that steers the next tool call is enough to reach it. Six layers, each one utility's state-changing form: (1) sort with -o / an -o short-option cluster / --output (and its unique prefixes --o, --out ...), or --compress-program naming anything other than a real compressor (gzip, pigz, bzip2, xz, lz4, zstd, ...): sort executes that program, which turns a read-only allow-list entry into code execution; (2) diff with --output / --out / -o; (3) uniq with two non-option operands, where the second is the output file; option arguments of -f / -s / -w are consumed so 'uniq -f 2 data.txt' is not read as two operands; (4) date with -s / --set (and --se, --s) or a bare operand that is a valid MMDDhhmm[[CC]YY][.ss] timestamp (month 01-12, day 01-31, hour 00-23, minute 00-59), both of which set the system clock; a YYYYMMDD string such as 'date 20261022' is not a valid month and does not match; (5) hostname given a new name as its only operand, or -F / --file; (6) file with --compile (or its unique prefixes --co, --comp ...), which writes a compiled magic database (.mgc) into the working directory. Scope and limits: the utility must be the FIRST token of a JSON string value whose key is command-like: command, cmd, cmd_line, command_line, cmdline, commandline, shell, script, exec, args, argv, optionally prefixed by a word ending in '_' or '-' (bash_command, shell_cmd) or by bash, shell, sh, run, exec, raw or full (shellCommand, runCmd). Keys that merely end in those letters (transcript, subscript) do not count. The utility may be prefixed by sudo or a /usr/bin, /bin or ./ path. A JSON-escaped tab (\t) between tokens is treated as whitespace, because the shell splits on it; an escaped newline is not, because it ends the command. Commands reached after a pipe, ';', '&&' or a newline, passed as an argv array, or wrapped by env / nice / timeout are out of scope (shell-metacharacter rules cover the chained forms). file -C cannot be told apart from the read-only file -c because the engine matches case-insensitively, so only the long --compile form is covered. file -z / --uncompress, which the same fix also rejects, is not covered: it runs only libmagic's fixed decompressor list, not an attacker-chosen program, and 'file -z archive.gz' is ordinary read-only use. uniq operands that are quoted are not matched. CWE-88, CWE-184. RE2 portability: every condition compiles under RE2 (checked with Go regexp). The earlier lookarounds were replaced without changing behaviour: trailing lookaheads now consume the terminator, the uniq option/operand lookaheads became explicit alternatives, and the --compress-program compressor exception became a positive witness generated from the compressor/path trie. Measured equivalent to the lookaround version on 2.4M generated command strings and an exhaustive 1.48M-string sweep of the --compress-program value forms. The trade-off is size: Layer 1 is about 2.4 KB because a finite exception list has no short RE2 spelling. Known false positive: in a general-purpose shell tool where writing files is allowed, the ordinary write forms of these utilities match exactly like the attack, for example 'sort -o sorted.txt data.txt', in-place sorting with 'sort -u -o names.txt names.txt', and 'uniq -c sorted.txt counts.txt'. The rule reports that a read-only-looking utility was used to write, not intent, which is why severity is medium and the response is observe-only.

Response Actions

References

1. 01

   sort invoked with an output file (-o / --output) or with --compress-program naming a non-compressor program - sort writes or executes it itself (CVE-2026-105789 primary primitive)

   field: contentop: regex
2. 02

   Same as Layer 1, in the JSON-encoded tool argument object

   field: tool_argsop: regex
3. 03

   diff invoked with an output file (--output / --out / -o) - diff writes the destination itself

   field: contentop: regex
4. 04

   Same as Layer 2, in the JSON-encoded tool argument object

   field: tool_argsop: regex
5. 05

   uniq given an input and an output operand - uniq creates or overwrites the second file itself

   field: contentop: regex
6. 06

   Same as Layer 3, in the JSON-encoded tool argument object

   field: tool_argsop: regex
7. 07

   date invoked with -s / --set or a valid MMDDhhmm operand - both set the system clock

   field: contentop: regex
8. 08

   Same as Layer 4, in the JSON-encoded tool argument object

   field: tool_argsop: regex
9. 09

   hostname given a new name or a name file - hostname changes the system hostname

   field: contentop: regex
10. 10

    Same as Layer 5, in the JSON-encoded tool argument object

    field: tool_argsop: regex
11. 11

    file invoked with --compile - file writes a compiled magic database into the working directory

    field: contentop: regex
12. 12

    Same as Layer 6, in the JSON-encoded tool argument object

    field: tool_argsop: regex

Attack Examples (Rule Triggers)

1. Matches: Layer 1

   {"command":"sort -o /home/ufo/UFO/config/ufo/agents.yaml /tmp/junk.txt"}

2. Matches: Layer 1

   {"command":"sort --output=/home/ufo/.bashrc data.txt"}

3. Matches: Layer 1

   {"command":"sort -S 1K --compress-program=make data.txt"}

4. Matches: Layer 1

   {"tool":"run_shell","arguments":{"cmd":"sort -u -o ~/.config/app/settings.json notes.txt"}}

5. Matches: Layer 3

   {"command":"uniq input.txt /home/ufo/UFO/config/ufo/system.yaml"}

6. Matches: Layer 3

   {"command":"uniq -c input.txt output.txt"}

7. Matches: Layer 2

   {"command":"diff --output=result.txt original.txt updated.txt"}

8. Matches: Layer 2

   {"command":"diff -oresult.txt original.txt updated.txt"}

9. Matches: Layer 4

   {"command":"date -s 2030-01-01"}

10. Matches: Layer 4

    {"command":"date --se=2030-01-01"}

11. Matches: Layer 5

    {"command":"hostname attacker-host"}

12. Matches: Layer 4

    {"command":"date 090212002026"}

13. Matches: Layer 1

    {"cmd_line":"sort -o /home/ufo/.profile notes.txt"}

14. Matches: Layer 3

    {"command_line":"uniq data.txt /home/ufo/UFO/config/ufo/agents.yaml"}

15. Matches: Layer 1

    {"command":"sort\t-o\t/home/ufo/.bashrc\tdata.txt"}

16. Matches: Layer 6

    {"command":"file -C -m /home/ufo/magic --compile"}

17. Matches: Layer 6

    {"bash_command":"file --comp -m rules.magic"}

18. Matches: Layer 1

    {"command":"sort -S 1K --compress-program='/tmp/payload.sh' data.txt"}

19. Matches: Layer 1

    {"command":"sort --compress-program=/usr/bin/gzipx data.txt"}

Real-world attack payloads, sanitized and versioned alongside the rule as regression tests — so a future revision can't silently stop catching them.

Benign Examples (Rule Doesn't Trigger)

1. Benign twin: sort reading and printing, the read-only use the allow-list intended

   {"command":"sort -n data.txt"}

2. Benign twin: uniq with one operand, allowed by the 3.0.9 policy

   {"command":"uniq -c data.txt"}

3. Benign twin: -f takes the number, so there is still only one operand

   {"command":"uniq -f 2 -c access.log"}

4. Benign twin: diff printing to stdout, allowed by the 3.0.9 policy

   {"command":"diff original.txt updated.txt"}

5. Benign twin: diff long options that also start with --o

   {"command":"diff --old-line-format=x --new-line-format=y a.txt b.txt"}

6. Benign twin: reading the clock, allowed by the 3.0.9 policy

   {"command":"date -u +%s"}

7. Benign twin: date -d formats a date without setting the clock

   {"command":"date -d yesterday +%Y-%m-%d"}

8. Benign twin: reading the FQDN, allowed by the 3.0.9 policy

   {"command":"hostname -f"}

9. Benign twin: --compress-program with a real compressor for temp files

   {"command":"sort -k2,2 -t, -S 1G --compress-program=zstd big.csv"}

10. Benign twin: -t takes the letter o as the field separator; it is not -o

    {"command":"sort -to -k2 names.txt"}

11. Benign twin: classic read-only pipeline

    {"command":"sort data.txt | uniq -c"}

12. Search query, not a command-like key

    {"query":"sort -o output.txt input.txt"}

13. Advisory prose quoting the payload; no command value starts with the utility

    Before 3.0.9, execute_command allowed sort -o /home/ufo/.bashrc data.txt and the second uniq operand to write files.

14. Reviewer FP: transcript only ends in the letters 'script'; not a command-like key

    {"transcript":"date 20261022 standup notes: reviewed the deploy plan"}

15. Reviewer FP: subscript only ends in the letters 'script'; not a command-like key

    {"subscript":"hostname server01"}

16. Reviewer FP: a YYYYMMDD string is not a valid MMDDhhmm (month 20), date fails without touching the clock

    {"command":"date 20261022"}

17. Benign twin: file -z looks inside a compressed file with libmagic's fixed decompressors; out of scope

    {"command":"file -z archive.tar.gz"}

18. Benign twin: file -c (checking printout) is read-only

    {"command":"file -c -m /usr/share/misc/magic"}

19. Benign twin: an escaped newline ends the sort command, so it is not read as whitespace

    {"command":"sort data.txt\nuniq -c out.txt"}

20. RE2 rewrite check: real compressor given with a /usr/bin/ path stays excluded

    {"command":"sort -S 1G --compress-program=/usr/bin/pigz big.csv"}

21. RE2 rewrite check: -w takes the number, so only one operand remains

    {"command":"uniq -w 10 access.log"}

Known False Positive Contexts

• ▸Known false positive: a general-purpose shell tool where the user genuinely asks the agent to write sorted or de-duplicated output to a file produces the same tool call as the attack, for example sort -o sorted.txt data.txt, in-place sort -u -o names.txt names.txt, or uniq -c sorted.txt counts.txt. The rule cannot tell intent; it reports that a utility commonly allow-listed as read-only was used in its write form.
• ▸Administrative agents that are meant to set the clock or hostname (date -s, hostname new-name) will match.
• ▸Building a custom magic database with file -C -m magic --compile in a development shell will match.
• ▸Security test suites and CVE reproductions that pass these command strings to a tool will match, which is intended for a detection corpus.

Documented Evasion Techniques

1. Technique: chained command

   {"command":"cat data.txt && sort -o /home/ufo/.bashrc data.txt"}

   Out of scope: the utility is not the first token. Chained commands are the shell-metacharacter rules' territory.
2. Technique: argv array

   {"argv":["sort","-o","/home/ufo/.bashrc","data.txt"]}

   Real gap, accepted: argv arrays split the utility and its option into separate JSON strings.
3. Technique: wrapper binary

   {"command":"env sort -o /home/ufo/.bashrc data.txt"}

   Real gap, accepted: wrappers other than sudo (env, nice, timeout) are not stripped.
4. Technique: json escaped whitespace

   {"command":"sort -o\t/home/ufo/.bashrc data.txt"}

   Covered: a JSON-escaped tab is treated as whitespace. An escaped newline is deliberately not, since it terminates the command.
5. Technique: fixed decompressor

   {"command":"file -z /tmp/payload.gz"}

   Accepted: the fix also rejects file -z / --uncompress, but it only runs libmagic's fixed decompressors on the input and is common read-only use.

Publicly documented bypasses. A standard earns trust by publishing its worst figures, not hiding them — so known limitations ship inside the rule, not in a footnote.

title: "Read-Only Allow-Listed Utility Invoked With a File-Write or State-Changing Option in a Command Argument"
id: ATR-2026-02865
rule_version: 1
status: experimental
description: >
  Detects an agent or MCP tool call whose command-string argument starts with a
  utility that command allow-lists commonly classify as read-only (sort, uniq,
  diff, date, hostname, file) and gives it the option or operand that makes the
  utility itself write a file, run a helper program, or change system state.
  No shell metacharacter is involved: the allowed binary opens the destination
  on its own, so filters that only look for ; | && $( ) never see it.

  This is the shape reported as CVE-2026-105789 / GHSA-85w2-wggf-rw49 against
  Microsoft UFO before 3.0.9. The execute_command tool of the Linux MCP server
  (ufo/client/mcp/http_servers/linux_mcp_server.py) accepted a free-form
  command string and allowed sort and uniq as read-only commands, so
  'sort -o <file>' or the optional second uniq operand created or overwrote any
  file writable by the server process (agent configuration, shell startup
  files). The fix (commit 4f79362) removed sort and added per-command argument
  policies; its regression tests (tests/test_linux_mcp_command_policy.py) list
  the rejected forms used here. The command string is produced by the model,
  so a prompt injection that steers the next tool call is enough to reach it.

  Six layers, each one utility's state-changing form:
  (1) sort with -o / an -o short-option cluster / --output (and its unique
  prefixes --o, --out ...), or --compress-program naming anything other than a
  real compressor (gzip, pigz, bzip2, xz, lz4, zstd, ...): sort executes that
  program, which turns a read-only allow-list entry into code execution;
  (2) diff with --output / --out / -o;
  (3) uniq with two non-option operands, where the second is the output file;
  option arguments of -f / -s / -w are consumed so 'uniq -f 2 data.txt' is not
  read as two operands;
  (4) date with -s / --set (and --se, --s) or a bare operand that is a valid
  MMDDhhmm[[CC]YY][.ss] timestamp (month 01-12, day 01-31, hour 00-23, minute
  00-59), both of which set the system clock; a YYYYMMDD string such as
  'date 20261022' is not a valid month and does not match;
  (5) hostname given a new name as its only operand, or -F / --file;
  (6) file with --compile (or its unique prefixes --co, --comp ...), which
  writes a compiled magic database (.mgc) into the working directory.

  Scope and limits: the utility must be the FIRST token of a JSON string value
  whose key is command-like: command, cmd, cmd_line, command_line, cmdline,
  commandline, shell, script, exec, args, argv, optionally prefixed by a word
  ending in '_' or '-' (bash_command, shell_cmd) or by bash, shell, sh, run,
  exec, raw or full (shellCommand, runCmd). Keys that merely end in those
  letters (transcript, subscript) do not count. The utility may be prefixed by
  sudo or a /usr/bin, /bin or ./ path. A JSON-escaped tab (\t) between tokens is
  treated as whitespace, because the shell splits on it; an escaped newline is
  not, because it ends the command. Commands reached after a pipe, ';', '&&' or
  a newline, passed as an argv array, or wrapped by env / nice / timeout are
  out of scope (shell-metacharacter rules cover the chained forms). file -C
  cannot be told apart from the read-only file -c because the engine matches
  case-insensitively, so only the long --compile form is covered. file -z /
  --uncompress, which the same fix also rejects, is not covered: it runs only
  libmagic's fixed decompressor list, not an attacker-chosen program, and
  'file -z archive.gz' is ordinary read-only use. uniq operands that are quoted
  are not matched. CWE-88, CWE-184.

  RE2 portability: every condition compiles under RE2 (checked with Go
  regexp). The earlier lookarounds were replaced without changing behaviour:
  trailing lookaheads now consume the terminator, the uniq option/operand
  lookaheads became explicit alternatives, and the --compress-program
  compressor exception became a positive witness generated from the
  compressor/path trie. Measured equivalent to the lookaround version on
  2.4M generated command strings and an exhaustive 1.48M-string sweep of the
  --compress-program value forms. The trade-off is size: Layer 1 is about
  2.4 KB because a finite exception list has no short RE2 spelling.

  Known false positive: in a general-purpose shell tool where writing files is
  allowed, the ordinary write forms of these utilities match exactly like the
  attack, for example 'sort -o sorted.txt data.txt', in-place sorting with
  'sort -u -o names.txt names.txt', and 'uniq -c sorted.txt counts.txt'. The
  rule reports that a read-only-looking utility was used to write, not intent,
  which is why severity is medium and the response is observe-only.
author: "ATR Community (CVE sync)"
date: "2026/10/22"
schema_version: "0.1"
detection_tier: pattern
maturity: test
severity: medium

references:
  cve:
    - "CVE-2026-105789"
  ghsa:
    - "GHSA-85w2-wggf-rw49"
  cwe:
    - "CWE-88"
    - "CWE-184"
  owasp_llm:
    - "LLM06:2025"
    - "LLM03:2025"
  owasp_agentic:
    - "ASI02:2026"
    - "ASI05:2026"
    - "ASI04:2026"
  mitre_atlas:
    - "AML.T0053 - AI Agent Tool Invocation"
  mitre_attack:
    - "T1565.001 - Data Manipulation: Stored Data Manipulation"
    - "T1546.004 - Event Triggered Execution: Unix Shell Configuration Modification"
    - "T1202 - Indirect Command Execution"
  external:
    - "https://github.com/microsoft/UFO/security/advisories/GHSA-85w2-wggf-rw49"
    - "https://github.com/microsoft/UFO/pull/349"
    - "https://github.com/microsoft/UFO/commit/4f793622794f5d47810d54bed431c61f6a781dd6"
    - "https://github.com/microsoft/UFO/releases/tag/v3.0.9"
    - "https://euvd.enisa.europa.eu/vulnerability/EUVD-2026-93406"

metadata_provenance:
  cve: euvd-sync
  cwe: human-reviewed
  owasp_llm: human-reviewed
  owasp_agentic: human-reviewed
  mitre_atlas: human-reviewed
  mitre_attack: human-reviewed

compliance:
  eu_ai_act:
    - article: "15"
      context: "Article 15 requires cybersecurity appropriate to the intended purpose. An agent whose command tool trusts sort, uniq, diff, date, hostname or file as read-only can be steered into overwriting its own configuration or changing system state; flagging those utilities' write and set forms provides detection evidence supporting Article 15 robustness at the agent-to-OS boundary."
      strength: primary
    - article: "9"
      context: "Article 9 risk management must identify how an agent's tools can be misused. Matches from this rule provide detection evidence supporting Article 9 by showing when a command allow-list's read-only classification was exceeded through a utility's own options, a risk that first-token allow-lists do not capture."
      strength: secondary
  nist_ai_rmf:
    - function: Measure
      subcategory: "MS.2.7"
      context: "MS.2.7 requires security and resilience to be evaluated and documented. Inspecting each command-tool invocation for file-write and state-changing options of allow-listed utilities is a measurable per-call control rather than an assumption that the allow-list is complete."
      strength: primary
    - function: Manage
      subcategory: "MG.3.2"
      context: "MG.3.2 covers monitoring of third-party components in operation. The vulnerable execute_command tool ships with an agent framework the deployer did not write; monitoring the commands it actually receives is the operational check on that component."
      strength: secondary
  iso_42001:
    - clause: "8.1"
      context: "Clause 8.1 operational control over externally provided processes covers the MCP servers an agent invokes. This rule is applied to every command those servers receive, independent of the server's own argument policy."
      strength: primary
    - clause: "8.3"
      context: "Clause 8.3 AI risk treatment requires controls proportionate to the risk. Because one sort -o can overwrite the agent's own configuration, runtime detection of the write form is a proportionate treatment alongside upgrading the framework."
      strength: secondary

tags:
  category: tool-poisoning
  subcategory: read-only-utility-write-option
  scan_target: tool_args
  confidence: medium
  source: cve-disclosure
  vendor_sources: microsoft-ufo-cve-2026-105789

agent_source:
  type: mcp_exchange
  framework:
    - mcp
    - any
  provider:
    - any

detection:
  condition: any
  method: pattern
  false_positives:
    - "Known false positive: a general-purpose shell tool where the user genuinely asks the agent to write sorted or de-duplicated output to a file produces the same tool call as the attack, for example sort -o sorted.txt data.txt, in-place sort -u -o names.txt names.txt, or uniq -c sorted.txt counts.txt. The rule cannot tell intent; it reports that a utility commonly allow-listed as read-only was used in its write form."
    - "Administrative agents that are meant to set the clock or hostname (date -s, hostname new-name) will match."
    - "Building a custom magic database with file -C -m magic --compile in a development shell will match."
    - "Security test suites and CVE reproductions that pass these command strings to a tool will match, which is intended for a detection corpus."
  conditions:
    # Shared pieces (repeated per layer because the engine has no macros; built by
    # a generator so the six layers stay identical). Every pattern is RE2-portable:
    # no lookahead / lookbehind, every repeat bound <= 1000, and nested counted
    # repeats multiply to <= 1000 (Go regexp rejects a larger product).
    #   KEY  \\{0,7}"(?:\w{0,24}?[_-]|bash|shell|sh|run|exec|raw|full)?(?:command|cmd|...)\\{0,7}"\s*:\s*\\{0,7}"
    #        command-like JSON key, quotes JSON-escaped up to 3 levels; a prefix must
    #        end in _ or - or be a listed word, so transcript / subscript do not count
    #   SEP  (?:\s|\\(?:\\\\?)?t)   whitespace or a JSON-escaped tab (1-3 backslashes);
    #        written with ? instead of {1,3} so it can nest inside {1,64} and {0,8}
    #   PFX  (?:sudo SEP{1,64})?(?:/(?:usr/)?(?:local/)?s?bin/|\./)?   optional sudo / path
    #   SKIP (?:SEP{1,64}[^\s"\\|;&<>]{1,80}){0,8}?SEP{1,64}   other tokens of the SAME
    #        command; cannot cross a quote, backslash or shell operator
    #   ARG  (?:\\{1,7}"|')?[^\s"\\|;&<>{}\[\],-]   the destination starts; an escaped
    #        quote may open it, but a closing quote followed by } , ] does not
    #   END  (?:SEP|\\{0,7}"|$)   consumes the token terminator (was a trailing lookahead;
    #        consuming it is equivalent for a yes/no match)
    # Every backslash run is bounded ({0,7} / {1,7} / {8}); an unbounded \\* before a
    # quote is quadratic on a long backslash run.
    #
    # Layer 1 - sort -o / --output / --compress-program=<non-compressor>
    # The compressor allow-list used to be a negative lookahead. It is now a
    # positive "witness" that the program token is NOT one of gzip, pigz, bzip2,
    # pbzip2, lbzip2, xz, pxz, lz4, lzop, zstd, pzstd, lzma, lzip, plzip, brotli
    # (optionally after /bin/, /sbin/, /usr/bin/, /usr/sbin/, /usr/local/bin/ ...)
    # followed by a terminator: either the token leaves that name/path trie at
    # some character, ends early, or a full name is followed by something that
    # is not whitespace, a quote, \\{0,7}" or \\{1,3}t. Generated from the trie.
    - field: content
      operator: regex
      value: '\\{0,7}"(?:\w{0,24}?[_-]|bash|shell|sh|run|exec|raw|full)?(?:command_line|commandline|cmd_line|cmdline|command|cmd|shell|script|exec|args?|argv)\\{0,7}"\s*:\s*\\{0,7}"\s{0,16}(?:sudo(?:\s|\\(?:\\\\?)?t){1,64})?(?:/(?:usr/)?(?:local/)?s?bin/|\./)?sort(?:(?:\s|\\(?:\\\\?)?t){1,64}[^\s"\\|;&<>]{1,80}){0,8}?(?:\s|\\(?:\\\\?)?t){1,64}(?:(?:-[bcdfghimnrsuvz]{0,8}o(?:\s|\\(?:\\\\?)?t){0,64}|--o(?:u(?:t(?:p(?:u(?:t)?)?)?)?)?(?:=|(?:\s|\\(?:\\\\?)?t){1,64}))(?:\\{1,7}"|'')?[^\s"\\|;&<>{}\[\],-]|--co[a-z-]{0,16}(?:=|(?:\s|\\(?:\\\\?)?t){1,64})(?:''(?:[^/bglpxz]|$)|\\{1,7}"[^\s"\\|;&<>{}\[\],/bglpxz-]|[^\s"\\|;&<>{}\[\],''/bglpxz-]|(?:''|\\{1,7}")?(?:/(?:usr/)?(?:local/)?s?bin/(?:[^bglpxz]|$|b(?:[^rz]|$|r(?:[^o]|$|o(?:[^t]|$|t(?:[^l]|$|l(?:[^i]|$))))|z(?:[^i]|$|i(?:[^p]|$|p(?:[^2]|$))))|g(?:[^z]|$|z(?:[^i]|$|i(?:[^p]|$)))|l(?:[^bz]|$|b(?:[^z]|$|z(?:[^i]|$|i(?:[^p]|$|p(?:[^2]|$))))|z(?:[^4imo]|$|i(?:[^p]|$)|m(?:[^a]|$)|o(?:[^p]|$)))|p(?:[^bilxz]|$|b(?:[^z]|$|z(?:[^i]|$|i(?:[^p]|$|p(?:[^2]|$))))|i(?:[^g]|$|g(?:[^z]|$))|l(?:[^z]|$|z(?:[^i]|$|i(?:[^p]|$)))|x(?:[^z]|$)|z(?:[^s]|$|s(?:[^t]|$|t(?:[^d]|$))))|x(?:[^z]|$)|z(?:[^s]|$|s(?:[^t]|$|t(?:[^d]|$)))|(?:gzip|pigz|bzip2|pbzip2|lbzip2|xz|pxz|lz4|lzop|zstd|pzstd|lzma|lzip|plzip|brotli)(?:[^\s''"\\]|\\{8}|\\{1,7}(?:[^"\\t]|$)|\\{4,7}t))|/(?:[^blsu]|$|b(?:[^i]|$|i(?:[^n]|$|n(?:[^/]|$)))|l(?:[^o]|$|o(?:[^c]|$|c(?:[^a]|$|a(?:[^l]|$|l(?:[^/]|$|/(?:[^bs]|$|b(?:[^i]|$|i(?:[^n]|$|n(?:[^/]|$)))|s(?:[^b]|$|b(?:[^i]|$|i(?:[^n]|$|n(?:[^/]|$))))))))))|s(?:[^b]|$|b(?:[^i]|$|i(?:[^n]|$|n(?:[^/]|$))))|u(?:[^s]|$|s(?:[^r]|$|r(?:[^/]|$|/(?:[^bls]|$|b(?:[^i]|$|i(?:[^n]|$|n(?:[^/]|$)))|l(?:[^o]|$|o(?:[^c]|$|c(?:[^a]|$|a(?:[^l]|$|l(?:[^/]|$|/(?:[^bs]|$|b(?:[^i]|$|i(?:[^n]|$|n(?:[^/]|$)))|s(?:[^b]|$|b(?:[^i]|$|i(?:[^n]|$|n(?:[^/]|$))))))))))|s(?:[^b]|$|b(?:[^i]|$|i(?:[^n]|$|n(?:[^/]|$)))))))))|(?:b(?:[^rz]|$|r(?:[^o]|$|o(?:[^t]|$|t(?:[^l]|$|l(?:[^i]|$))))|z(?:[^i]|$|i(?:[^p]|$|p(?:[^2]|$))))|g(?:[^z]|$|z(?:[^i]|$|i(?:[^p]|$)))|l(?:[^bz]|$|b(?:[^z]|$|z(?:[^i]|$|i(?:[^p]|$|p(?:[^2]|$))))|z(?:[^4imo]|$|i(?:[^p]|$)|m(?:[^a]|$)|o(?:[^p]|$)))|p(?:[^bilxz]|$|b(?:[^z]|$|z(?:[^i]|$|i(?:[^p]|$|p(?:[^2]|$))))|i(?:[^g]|$|g(?:[^z]|$))|l(?:[^z]|$|z(?:[^i]|$|i(?:[^p]|$)))|x(?:[^z]|$)|z(?:[^s]|$|s(?:[^t]|$|t(?:[^d]|$))))|x(?:[^z]|$)|z(?:[^s]|$|s(?:[^t]|$|t(?:[^d]|$)))|(?:gzip|pigz|bzip2|pbzip2|lbzip2|xz|pxz|lz4|lzop|zstd|pzstd|lzma|lzip|plzip|brotli)(?:[^\s''"\\]|\\{8}|\\{1,7}(?:[^"\\t]|$)|\\{4,7}t)))))'
      description: "sort invoked with an output file (-o / --output) or with --compress-program naming a non-compressor program - sort writes or executes it itself (CVE-2026-105789 primary primitive)"
    - field: tool_args
      operator: regex
      value: '\\{0,7}"(?:\w{0,24}?[_-]|bash|shell|sh|run|exec|raw|full)?(?:command_line|commandline|cmd_line|cmdline|command|cmd|shell|script|exec|args?|argv)\\{0,7}"\s*:\s*\\{0,7}"\s{0,16}(?:sudo(?:\s|\\(?:\\\\?)?t){1,64})?(?:/(?:usr/)?(?:local/)?s?bin/|\./)?sort(?:(?:\s|\\(?:\\\\?)?t){1,64}[^\s"\\|;&<>]{1,80}){0,8}?(?:\s|\\(?:\\\\?)?t){1,64}(?:(?:-[bcdfghimnrsuvz]{0,8}o(?:\s|\\(?:\\\\?)?t){0,64}|--o(?:u(?:t(?:p(?:u(?:t)?)?)?)?)?(?:=|(?:\s|\\(?:\\\\?)?t){1,64}))(?:\\{1,7}"|'')?[^\s"\\|;&<>{}\[\],-]|--co[a-z-]{0,16}(?:=|(?:\s|\\(?:\\\\?)?t){1,64})(?:''(?:[^/bglpxz]|$)|\\{1,7}"[^\s"\\|;&<>{}\[\],/bglpxz-]|[^\s"\\|;&<>{}\[\],''/bglpxz-]|(?:''|\\{1,7}")?(?:/(?:usr/)?(?:local/)?s?bin/(?:[^bglpxz]|$|b(?:[^rz]|$|r(?:[^o]|$|o(?:[^t]|$|t(?:[^l]|$|l(?:[^i]|$))))|z(?:[^i]|$|i(?:[^p]|$|p(?:[^2]|$))))|g(?:[^z]|$|z(?:[^i]|$|i(?:[^p]|$)))|l(?:[^bz]|$|b(?:[^z]|$|z(?:[^i]|$|i(?:[^p]|$|p(?:[^2]|$))))|z(?:[^4imo]|$|i(?:[^p]|$)|m(?:[^a]|$)|o(?:[^p]|$)))|p(?:[^bilxz]|$|b(?:[^z]|$|z(?:[^i]|$|i(?:[^p]|$|p(?:[^2]|$))))|i(?:[^g]|$|g(?:[^z]|$))|l(?:[^z]|$|z(?:[^i]|$|i(?:[^p]|$)))|x(?:[^z]|$)|z(?:[^s]|$|s(?:[^t]|$|t(?:[^d]|$))))|x(?:[^z]|$)|z(?:[^s]|$|s(?:[^t]|$|t(?:[^d]|$)))|(?:gzip|pigz|bzip2|pbzip2|lbzip2|xz|pxz|lz4|lzop|zstd|pzstd|lzma|lzip|plzip|brotli)(?:[^\s''"\\]|\\{8}|\\{1,7}(?:[^"\\t]|$)|\\{4,7}t))|/(?:[^blsu]|$|b(?:[^i]|$|i(?:[^n]|$|n(?:[^/]|$)))|l(?:[^o]|$|o(?:[^c]|$|c(?:[^a]|$|a(?:[^l]|$|l(?:[^/]|$|/(?:[^bs]|$|b(?:[^i]|$|i(?:[^n]|$|n(?:[^/]|$)))|s(?:[^b]|$|b(?:[^i]|$|i(?:[^n]|$|n(?:[^/]|$))))))))))|s(?:[^b]|$|b(?:[^i]|$|i(?:[^n]|$|n(?:[^/]|$))))|u(?:[^s]|$|s(?:[^r]|$|r(?:[^/]|$|/(?:[^bls]|$|b(?:[^i]|$|i(?:[^n]|$|n(?:[^/]|$)))|l(?:[^o]|$|o(?:[^c]|$|c(?:[^a]|$|a(?:[^l]|$|l(?:[^/]|$|/(?:[^bs]|$|b(?:[^i]|$|i(?:[^n]|$|n(?:[^/]|$)))|s(?:[^b]|$|b(?:[^i]|$|i(?:[^n]|$|n(?:[^/]|$))))))))))|s(?:[^b]|$|b(?:[^i]|$|i(?:[^n]|$|n(?:[^/]|$)))))))))|(?:b(?:[^rz]|$|r(?:[^o]|$|o(?:[^t]|$|t(?:[^l]|$|l(?:[^i]|$))))|z(?:[^i]|$|i(?:[^p]|$|p(?:[^2]|$))))|g(?:[^z]|$|z(?:[^i]|$|i(?:[^p]|$)))|l(?:[^bz]|$|b(?:[^z]|$|z(?:[^i]|$|i(?:[^p]|$|p(?:[^2]|$))))|z(?:[^4imo]|$|i(?:[^p]|$)|m(?:[^a]|$)|o(?:[^p]|$)))|p(?:[^bilxz]|$|b(?:[^z]|$|z(?:[^i]|$|i(?:[^p]|$|p(?:[^2]|$))))|i(?:[^g]|$|g(?:[^z]|$))|l(?:[^z]|$|z(?:[^i]|$|i(?:[^p]|$)))|x(?:[^z]|$)|z(?:[^s]|$|s(?:[^t]|$|t(?:[^d]|$))))|x(?:[^z]|$)|z(?:[^s]|$|s(?:[^t]|$|t(?:[^d]|$)))|(?:gzip|pigz|bzip2|pbzip2|lbzip2|xz|pxz|lz4|lzop|zstd|pzstd|lzma|lzip|plzip|brotli)(?:[^\s''"\\]|\\{8}|\\{1,7}(?:[^"\\t]|$)|\\{4,7}t)))))'
      description: "Same as Layer 1, in the JSON-encoded tool argument object"

    # Layer 2 - diff --output / --out / -o
    - field: content
      operator: regex
      value: '\\{0,7}"(?:\w{0,24}?[_-]|bash|shell|sh|run|exec|raw|full)?(?:command_line|commandline|cmd_line|cmdline|command|cmd|shell|script|exec|args?|argv)\\{0,7}"\s*:\s*\\{0,7}"\s{0,16}(?:sudo(?:\s|\\(?:\\\\?)?t){1,64})?(?:/(?:usr/)?(?:local/)?s?bin/|\./)?diff(?:(?:\s|\\(?:\\\\?)?t){1,64}[^\s"\\|;&<>]{1,80}){0,8}?(?:\s|\\(?:\\\\?)?t){1,64}(?:-[abcehnpqrtuvwyz]{0,6}o(?:\s|\\(?:\\\\?)?t){0,64}|--out(?:p(?:u(?:t)?)?)?(?:=|(?:\s|\\(?:\\\\?)?t){1,64}))(?:\\{1,7}"|'')?[^\s"\\|;&<>{}\[\],-]'
      description: "diff invoked with an output file (--output / --out / -o) - diff writes the destination itself"
    - field: tool_args
      operator: regex
      value: '\\{0,7}"(?:\w{0,24}?[_-]|bash|shell|sh|run|exec|raw|full)?(?:command_line|commandline|cmd_line|cmdline|command|cmd|shell|script|exec|args?|argv)\\{0,7}"\s*:\s*\\{0,7}"\s{0,16}(?:sudo(?:\s|\\(?:\\\\?)?t){1,64})?(?:/(?:usr/)?(?:local/)?s?bin/|\./)?diff(?:(?:\s|\\(?:\\\\?)?t){1,64}[^\s"\\|;&<>]{1,80}){0,8}?(?:\s|\\(?:\\\\?)?t){1,64}(?:-[abcehnpqrtuvwyz]{0,6}o(?:\s|\\(?:\\\\?)?t){0,64}|--out(?:p(?:u(?:t)?)?)?(?:=|(?:\s|\\(?:\\\\?)?t){1,64}))(?:\\{1,7}"|'')?[^\s"\\|;&<>{}\[\],-]'
      description: "Same as Layer 2, in the JSON-encoded tool argument object"

    # Layer 3 - uniq with a second (output) operand. Options are consumed first;
    # -f/-s/-w style options that are followed by a number must take it, so the
    # number is never counted as an operand. Without lookaround this is spelled
    # as: no options, or up to five options followed by a last option that is
    # either argument-less (then the first operand may not start with a digit)
    # or ends in its argument / -- (then it may).
    - field: content
      operator: regex
      value: '\\{0,7}"(?:\w{0,24}?[_-]|bash|shell|sh|run|exec|raw|full)?(?:command_line|commandline|cmd_line|cmdline|command|cmd|shell|script|exec|args?|argv)\\{0,7}"\s*:\s*\\{0,7}"\s{0,16}(?:sudo(?:\s|\\(?:\\\\?)?t){1,64})?(?:/(?:usr/)?(?:local/)?s?bin/|\./)?uniq(?:(?:\s|\\(?:\\\\?)?t){1,64}[^\s"\\|;&<>{}\[\],-][^\s"\\|;&<>]{0,200}|(?:(?:\s|\\(?:\\\\?)?t){1,64}(?:--|-[a-z\d]{1,6}(?:(?:\s|\\(?:\\\\?)?t){1,64}\d{1,6})?|--[a-z][a-z-]{1,20}(?:=[\w-]{0,20}|(?:\s|\\(?:\\\\?)?t){1,64}\d{1,6})?)){0,5}(?:\s|\\(?:\\\\?)?t){1,64}(?:(?:-[a-z\d]{1,6}|--[a-z][a-z-]{1,20})(?:\s|\\(?:\\\\?)?t){1,64}[^\s"\\|;&<>{}\[\],\d-][^\s"\\|;&<>]{0,200}|(?:--|-[a-z\d]{1,6}(?:\s|\\(?:\\\\?)?t){1,64}\d{1,6}|--[a-z][a-z-]{1,20}(?:=[\w-]{0,20}|(?:\s|\\(?:\\\\?)?t){1,64}\d{1,6}))(?:\s|\\(?:\\\\?)?t){1,64}[^\s"\\|;&<>{}\[\],-][^\s"\\|;&<>]{0,200}))(?:\s|\\(?:\\\\?)?t){1,64}[^\s"\\|;&<>{}\[\],-][^\s"\\|;&<>]{0,200}(?:(?:\s|\\(?:\\\\?)?t)|\\{0,7}"|$)'
      description: "uniq given an input and an output operand - uniq creates or overwrites the second file itself"
    - field: tool_args
      operator: regex
      value: '\\{0,7}"(?:\w{0,24}?[_-]|bash|shell|sh|run|exec|raw|full)?(?:command_line|commandline|cmd_line|cmdline|command|cmd|shell|script|exec|args?|argv)\\{0,7}"\s*:\s*\\{0,7}"\s{0,16}(?:sudo(?:\s|\\(?:\\\\?)?t){1,64})?(?:/(?:usr/)?(?:local/)?s?bin/|\./)?uniq(?:(?:\s|\\(?:\\\\?)?t){1,64}[^\s"\\|;&<>{}\[\],-][^\s"\\|;&<>]{0,200}|(?:(?:\s|\\(?:\\\\?)?t){1,64}(?:--|-[a-z\d]{1,6}(?:(?:\s|\\(?:\\\\?)?t){1,64}\d{1,6})?|--[a-z][a-z-]{1,20}(?:=[\w-]{0,20}|(?:\s|\\(?:\\\\?)?t){1,64}\d{1,6})?)){0,5}(?:\s|\\(?:\\\\?)?t){1,64}(?:(?:-[a-z\d]{1,6}|--[a-z][a-z-]{1,20})(?:\s|\\(?:\\\\?)?t){1,64}[^\s"\\|;&<>{}\[\],\d-][^\s"\\|;&<>]{0,200}|(?:--|-[a-z\d]{1,6}(?:\s|\\(?:\\\\?)?t){1,64}\d{1,6}|--[a-z][a-z-]{1,20}(?:=[\w-]{0,20}|(?:\s|\\(?:\\\\?)?t){1,64}\d{1,6}))(?:\s|\\(?:\\\\?)?t){1,64}[^\s"\\|;&<>{}\[\],-][^\s"\\|;&<>]{0,200}))(?:\s|\\(?:\\\\?)?t){1,64}[^\s"\\|;&<>{}\[\],-][^\s"\\|;&<>]{0,200}(?:(?:\s|\\(?:\\\\?)?t)|\\{0,7}"|$)'
      description: "Same as Layer 3, in the JSON-encoded tool argument object"

    # Layer 4 - date -s / --set / a valid MMDDhhmm[[CC]YY][.ss] operand
    - field: content
      operator: regex
      value: '\\{0,7}"(?:\w{0,24}?[_-]|bash|shell|sh|run|exec|raw|full)?(?:command_line|commandline|cmd_line|cmdline|command|cmd|shell|script|exec|args?|argv)\\{0,7}"\s*:\s*\\{0,7}"\s{0,16}(?:sudo(?:\s|\\(?:\\\\?)?t){1,64})?(?:/(?:usr/)?(?:local/)?s?bin/|\./)?date(?:(?:(?:\s|\\(?:\\\\?)?t){1,64}[^\s"\\|;&<>]{1,80}){0,8}?(?:\s|\\(?:\\\\?)?t){1,64}(?:-u?s(?:\s|\\(?:\\\\?)?t){0,64}|--s(?:et?)?(?:=|(?:\s|\\(?:\\\\?)?t){1,64}))(?:\\{1,7}"|'')?[^\s"\\|;&<>{}\[\],-]|(?:(?:\s|\\(?:\\\\?)?t){1,64}(?:-u|--utc|--universal)){0,3}(?:\s|\\(?:\\\\?)?t){1,64}(?:0[1-9]|1[0-2])(?:0[1-9]|[12]\d|3[01])(?:[01]\d|2[0-3])[0-5]\d(?:\d{2}){0,2}(?:\.\d{2})?(?:(?:\s|\\(?:\\\\?)?t)|\\{0,7}"|$))'
      description: "date invoked with -s / --set or a valid MMDDhhmm operand - both set the system clock"
    - field: tool_args
      operator: regex
      value: '\\{0,7}"(?:\w{0,24}?[_-]|bash|shell|sh|run|exec|raw|full)?(?:command_line|commandline|cmd_line|cmdline|command|cmd|shell|script|exec|args?|argv)\\{0,7}"\s*:\s*\\{0,7}"\s{0,16}(?:sudo(?:\s|\\(?:\\\\?)?t){1,64})?(?:/(?:usr/)?(?:local/)?s?bin/|\./)?date(?:(?:(?:\s|\\(?:\\\\?)?t){1,64}[^\s"\\|;&<>]{1,80}){0,8}?(?:\s|\\(?:\\\\?)?t){1,64}(?:-u?s(?:\s|\\(?:\\\\?)?t){0,64}|--s(?:et?)?(?:=|(?:\s|\\(?:\\\\?)?t){1,64}))(?:\\{1,7}"|'')?[^\s"\\|;&<>{}\[\],-]|(?:(?:\s|\\(?:\\\\?)?t){1,64}(?:-u|--utc|--universal)){0,3}(?:\s|\\(?:\\\\?)?t){1,64}(?:0[1-9]|1[0-2])(?:0[1-9]|[12]\d|3[01])(?:[01]\d|2[0-3])[0-5]\d(?:\d{2}){0,2}(?:\.\d{2})?(?:(?:\s|\\(?:\\\\?)?t)|\\{0,7}"|$))'
      description: "Same as Layer 4, in the JSON-encoded tool argument object"

    # Layer 5 - hostname <new-name> as the only operand, or hostname -F / --file
    - field: content
      operator: regex
      value: '\\{0,7}"(?:\w{0,24}?[_-]|bash|shell|sh|run|exec|raw|full)?(?:command_line|commandline|cmd_line|cmdline|command|cmd|shell|script|exec|args?|argv)\\{0,7}"\s*:\s*\\{0,7}"\s{0,16}(?:sudo(?:\s|\\(?:\\\\?)?t){1,64})?(?:/(?:usr/)?(?:local/)?s?bin/|\./)?hostname(?:(?:\s|\\(?:\\\\?)?t){1,64}(?:-[bv]{1,2}|--boot|--verbose)){0,3}(?:\s|\\(?:\\\\?)?t){1,64}(?:(?:-[bv]{0,2}f(?:\s|\\(?:\\\\?)?t){0,64}|--fi(?:le?)?(?:=|(?:\s|\\(?:\\\\?)?t){1,64}))(?:\\{1,7}"|'')?[^\s"\\|;&<>{}\[\],-]|[a-z0-9][a-z0-9.-]{0,252}(?:\s|\\(?:\\\\?)?t){0,16}(?:\\{0,7}"|$))'
      description: "hostname given a new name or a name file - hostname changes the system hostname"
    - field: tool_args
      operator: regex
      value: '\\{0,7}"(?:\w{0,24}?[_-]|bash|shell|sh|run|exec|raw|full)?(?:command_line|commandline|cmd_line|cmdline|command|cmd|shell|script|exec|args?|argv)\\{0,7}"\s*:\s*\\{0,7}"\s{0,16}(?:sudo(?:\s|\\(?:\\\\?)?t){1,64})?(?:/(?:usr/)?(?:local/)?s?bin/|\./)?hostname(?:(?:\s|\\(?:\\\\?)?t){1,64}(?:-[bv]{1,2}|--boot|--verbose)){0,3}(?:\s|\\(?:\\\\?)?t){1,64}(?:(?:-[bv]{0,2}f(?:\s|\\(?:\\\\?)?t){0,64}|--fi(?:le?)?(?:=|(?:\s|\\(?:\\\\?)?t){1,64}))(?:\\{1,7}"|'')?[^\s"\\|;&<>{}\[\],-]|[a-z0-9][a-z0-9.-]{0,252}(?:\s|\\(?:\\\\?)?t){0,16}(?:\\{0,7}"|$))'
      description: "Same as Layer 5, in the JSON-encoded tool argument object"

    # Layer 6 - file --compile (and unique prefixes): writes a compiled .mgc magic file
    - field: content
      operator: regex
      value: '\\{0,7}"(?:\w{0,24}?[_-]|bash|shell|sh|run|exec|raw|full)?(?:command_line|commandline|cmd_line|cmdline|command|cmd|shell|script|exec|args?|argv)\\{0,7}"\s*:\s*\\{0,7}"\s{0,16}(?:sudo(?:\s|\\(?:\\\\?)?t){1,64})?(?:/(?:usr/)?(?:local/)?s?bin/|\./)?file(?:(?:\s|\\(?:\\\\?)?t){1,64}[^\s"\\|;&<>]{1,80}){0,8}?(?:\s|\\(?:\\\\?)?t){1,64}--co(?:m(?:p(?:i(?:l(?:e)?)?)?)?)?(?:(?:\s|\\(?:\\\\?)?t)|\\{0,7}"|$)'
      description: "file invoked with --compile - file writes a compiled magic database into the working directory"
    - field: tool_args
      operator: regex
      value: '\\{0,7}"(?:\w{0,24}?[_-]|bash|shell|sh|run|exec|raw|full)?(?:command_line|commandline|cmd_line|cmdline|command|cmd|shell|script|exec|args?|argv)\\{0,7}"\s*:\s*\\{0,7}"\s{0,16}(?:sudo(?:\s|\\(?:\\\\?)?t){1,64})?(?:/(?:usr/)?(?:local/)?s?bin/|\./)?file(?:(?:\s|\\(?:\\\\?)?t){1,64}[^\s"\\|;&<>]{1,80}){0,8}?(?:\s|\\(?:\\\\?)?t){1,64}--co(?:m(?:p(?:i(?:l(?:e)?)?)?)?)?(?:(?:\s|\\(?:\\\\?)?t)|\\{0,7}"|$)'
      description: "Same as Layer 6, in the JSON-encoded tool argument object"

response:
  actions:
    - alert
    - snapshot
  actions_rationale: >-
    response-action eligibility (docs/RESPONSE-ACTION-ELIGIBILITY.md): new rule with no
    FP measurement on the 65K benign corpus and a documented false positive on ordinary
    write forms in general-purpose shell tools; ceiling = observe. No block action.
  auto_response_threshold: high
  notify:
    - security_team
  message_template: >
    [ATR-2026-02865] MEDIUM: a command tool argument uses a utility that command
    allow-lists commonly treat as read-only (sort, uniq, diff, date, hostname,
    file) in a form that writes a file, runs a helper program or changes system
    state. This is the CVE-2026-105789 (Microsoft UFO execute_command) pattern.
    In a general-purpose shell tool this can be an ordinary write; check whether
    the tool is meant to be read-only and enforces a per-command argument
    policy, and upgrade UFO to 3.0.9 or later.

confidence: 60

test_cases:
  true_positives:
    - input: '{"command":"sort -o /home/ufo/UFO/config/ufo/agents.yaml /tmp/junk.txt"}'
      expected: triggered
      description: "Overwriting the UFO agent configuration through sort -o (advisory reproduction)"
      matched_condition: "Layer 1"
    - input: '{"command":"sort --output=/home/ufo/.bashrc data.txt"}'
      expected: triggered
      description: "sort --output into a shell startup file"
      matched_condition: "Layer 1"
    - input: '{"command":"sort -S 1K --compress-program=make data.txt"}'
      expected: triggered
      description: "From the fix tests: --compress-program makes sort execute an arbitrary program"
      matched_condition: "Layer 1"
    - input: '{"tool":"run_shell","arguments":{"cmd":"sort -u -o ~/.config/app/settings.json notes.txt"}}'
      expected: triggered
      description: "Derived variant: different tool and key, -o after another short option"
      matched_condition: "Layer 1"
    - input: '{"command":"uniq input.txt /home/ufo/UFO/config/ufo/system.yaml"}'
      expected: triggered
      description: "uniq second operand overwriting the UFO system configuration (advisory reproduction)"
      matched_condition: "Layer 3"
    - input: '{"command":"uniq -c input.txt output.txt"}'
      expected: triggered
      description: "From the fix tests: uniq with an option and an output operand"
      matched_condition: "Layer 3"
    - input: '{"command":"diff --output=result.txt original.txt updated.txt"}'
      expected: triggered
      description: "From the fix tests: diff --output"
      matched_condition: "Layer 2"
    - input: '{"command":"diff -oresult.txt original.txt updated.txt"}'
      expected: triggered
      description: "From the fix tests: diff -o with the file attached"
      matched_condition: "Layer 2"
    - input: '{"command":"date -s 2030-01-01"}'
      expected: triggered
      description: "From the fix tests: date -s sets the system clock"
      matched_condition: "Layer 4"
    - input: '{"command":"date --se=2030-01-01"}'
      expected: triggered
      description: "From the fix tests: abbreviated --set"
      matched_condition: "Layer 4"
    - input: '{"command":"hostname attacker-host"}'
      expected: triggered
      description: "From the fix tests: hostname given a new name"
      matched_condition: "Layer 5"
    - input: '{"command":"date 090212002026"}'
      expected: triggered
      description: "From the fix tests: a valid MMDDhhmmCCYY operand sets the clock (Sep 2 12:00 2026)"
      matched_condition: "Layer 4"
    - input: '{"cmd_line":"sort -o /home/ufo/.profile notes.txt"}'
      expected: triggered
      description: "Reviewer variant: cmd_line key, missed by round 1"
      matched_condition: "Layer 1"
    - input: '{"command_line":"uniq data.txt /home/ufo/UFO/config/ufo/agents.yaml"}'
      expected: triggered
      description: "Derived variant: command_line key with the uniq output operand"
      matched_condition: "Layer 3"
    - input: '{"command":"sort\t-o\t/home/ufo/.bashrc\tdata.txt"}'
      expected: triggered
      description: "Reviewer variant: JSON-escaped tabs between tokens; the shell splits on tab exactly as on space"
      matched_condition: "Layer 1"
    - input: '{"command":"file -C -m /home/ufo/magic --compile"}'
      expected: triggered
      description: "From the fix tests (file --compile family): file writes a compiled magic database"
      matched_condition: "Layer 6"
    - input: '{"bash_command":"file --comp -m rules.magic"}'
      expected: triggered
      description: "Derived variant: unique prefix --comp of --compile under a bash_command key"
      matched_condition: "Layer 6"
    - input: '{"command":"sort -S 1K --compress-program=''/tmp/payload.sh'' data.txt"}'
      expected: triggered
      description: "RE2 rewrite check: quoted non-compressor program for --compress-program still matches"
      matched_condition: "Layer 1"
    - input: '{"command":"sort --compress-program=/usr/bin/gzipx data.txt"}'
      expected: triggered
      description: "RE2 rewrite check: a name that only starts with a compressor name is not a compressor"
      matched_condition: "Layer 1"
  true_negatives:
    - input: '{"command":"sort -n data.txt"}'
      expected: not_triggered
      description: "Benign twin: sort reading and printing, the read-only use the allow-list intended"
    - input: '{"command":"uniq -c data.txt"}'
      expected: not_triggered
      description: "Benign twin: uniq with one operand, allowed by the 3.0.9 policy"
    - input: '{"command":"uniq -f 2 -c access.log"}'
      expected: not_triggered
      description: "Benign twin: -f takes the number, so there is still only one operand"
    - input: '{"command":"diff original.txt updated.txt"}'
      expected: not_triggered
      description: "Benign twin: diff printing to stdout, allowed by the 3.0.9 policy"
    - input: '{"command":"diff --old-line-format=x --new-line-format=y a.txt b.txt"}'
      expected: not_triggered
      description: "Benign twin: diff long options that also start with --o"
    - input: '{"command":"date -u +%s"}'
      expected: not_triggered
      description: "Benign twin: reading the clock, allowed by the 3.0.9 policy"
    - input: '{"command":"date -d yesterday +%Y-%m-%d"}'
      expected: not_triggered
      description: "Benign twin: date -d formats a date without setting the clock"
    - input: '{"command":"hostname -f"}'
      expected: not_triggered
      description: "Benign twin: reading the FQDN, allowed by the 3.0.9 policy"
    - input: '{"command":"sort -k2,2 -t, -S 1G --compress-program=zstd big.csv"}'
      expected: not_triggered
      description: "Benign twin: --compress-program with a real compressor for temp files"
    - input: '{"command":"sort -to -k2 names.txt"}'
      expected: not_triggered
      description: "Benign twin: -t takes the letter o as the field separator; it is not -o"
    - input: '{"command":"sort data.txt | uniq -c"}'
      expected: not_triggered
      description: "Benign twin: classic read-only pipeline"
    - input: '{"query":"sort -o output.txt input.txt"}'
      expected: not_triggered
      description: "Search query, not a command-like key"
    - input: 'Before 3.0.9, execute_command allowed sort -o /home/ufo/.bashrc data.txt and the second uniq operand to write files.'
      expected: not_triggered
      description: "Advisory prose quoting the payload; no command value starts with the utility"
    - input: '{"transcript":"date 20261022 standup notes: reviewed the deploy plan"}'
      expected: not_triggered
      description: "Reviewer FP: transcript only ends in the letters 'script'; not a command-like key"
    - input: '{"subscript":"hostname server01"}'
      expected: not_triggered
      description: "Reviewer FP: subscript only ends in the letters 'script'; not a command-like key"
    - input: '{"command":"date 20261022"}'
      expected: not_triggered
      description: "Reviewer FP: a YYYYMMDD string is not a valid MMDDhhmm (month 20), date fails without touching the clock"
    - input: '{"command":"file -z archive.tar.gz"}'
      expected: not_triggered
      description: "Benign twin: file -z looks inside a compressed file with libmagic's fixed decompressors; out of scope"
    - input: '{"command":"file -c -m /usr/share/misc/magic"}'
      expected: not_triggered
      description: "Benign twin: file -c (checking printout) is read-only"
    - input: '{"command":"sort data.txt\nuniq -c out.txt"}'
      expected: not_triggered
      description: "Benign twin: an escaped newline ends the sort command, so it is not read as whitespace"
    - input: '{"command":"sort -S 1G --compress-program=/usr/bin/pigz big.csv"}'
      expected: not_triggered
      description: "RE2 rewrite check: real compressor given with a /usr/bin/ path stays excluded"
    - input: '{"command":"uniq -w 10 access.log"}'
      expected: not_triggered
      description: "RE2 rewrite check: -w takes the number, so only one operand remains"

evasion_tests:
  - input: '{"command":"cat data.txt && sort -o /home/ufo/.bashrc data.txt"}'
    expected: not_triggered
    bypass_technique: chained_command
    notes: "Out of scope: the utility is not the first token. Chained commands are the shell-metacharacter rules' territory."
  - input: '{"argv":["sort","-o","/home/ufo/.bashrc","data.txt"]}'
    expected: not_triggered
    bypass_technique: argv_array
    notes: "Real gap, accepted: argv arrays split the utility and its option into separate JSON strings."
  - input: '{"command":"env sort -o /home/ufo/.bashrc data.txt"}'
    expected: not_triggered
    bypass_technique: wrapper_binary
    notes: "Real gap, accepted: wrappers other than sudo (env, nice, timeout) are not stripped."
  - input: '{"command":"sort -o\t/home/ufo/.bashrc data.txt"}'
    expected: triggered
    bypass_technique: json_escaped_whitespace
    notes: "Covered: a JSON-escaped tab is treated as whitespace. An escaped newline is deliberately not, since it terminates the command."
  - input: '{"command":"file -z /tmp/payload.gz"}'
    expected: not_triggered
    bypass_technique: fixed_decompressor
    notes: "Accepted: the fix also rejects file -z / --uncompress, but it only runs libmagic's fixed decompressors on the input and is common read-only use."