Import of a documented AI-hallucinated (slopsquatting) package name

SIGMA detection rule by WinstonRedGuard -- sigma plugin ai-fingerprint observed rules · Unknown

Detects an `import`/`from...import` statement naming a specific, independently-documented package name that LLM coding assistants are known to hallucinate (e.g. huggingface_cli -- the real package is huggingface_hub). Curated seed list, not a heuristic: each entry traces to a cited, confirmed-nonexistent-at-discovery package. A hit means the code names a package with no legitimate reason to exist, which is exactly the shape of a supply-chain compromise (slopsquatting: an attacker registers the invented name and ships malware to whoever installs it). Per-hit confidence is high without requiring co-occurrence, same reasoning as ai_provenance -- a literal match against a definitively-known-fake name is near-zero-FP by construction. Aggregated from 1 fingerprint hit(s) across 1 file(s). Sample paths: apps/foo/corge.py.

title: Import of a documented AI-hallucinated (slopsquatting) package name
id: ba9168d9-c764-56fa-9414-ecf3523d3d50
status: experimental
description: 'Detects an `import`/`from...import` statement naming a specific, independently-documented package name that
  LLM coding assistants are known to hallucinate (e.g. huggingface_cli -- the real package is huggingface_hub). Curated seed
  list, not a heuristic: each entry traces to a cited, confirmed-nonexistent-at-discovery package. A hit means the code names
  a package with no legitimate reason to exist, which is exactly the shape of a supply-chain compromise (slopsquatting: an
  attacker registers the invented name and ships malware to whoever installs it). Per-hit confidence is high without requiring
  co-occurrence, same reasoning as ai_provenance -- a literal match against a definitively-known-fake name is near-zero-FP
  by construction.


  Aggregated from 1 fingerprint hit(s) across 1 file(s). Sample paths: apps/foo/corge.py.'
references:
- https://cloud.google.com/blog/topics/threat-intelligence/ai-vulnerability-exploitation-initial-access
author: WinstonRedGuard -- sigma plugin ai-fingerprint observed rules
date: '2026-08-04'
logsource:
  category: code_review
  product: wrg_ai_fingerprint
detection:
  selection:
    detector: hallucinated_import
    evidence|contains:
    - 'hallucinated package import: ''huggingface_cli'''
  condition: selection
falsepositives:
- None expected for a listed name -- if it becomes a real, legitimately-registered package after the entry was added, remove
  the entry rather than suppress at the rule level
level: high
tags:
- attack.t1195.002
- wrg.ai_fingerprint
- wrg.detector.hallucinated_import
- wrg.severity.high
- wrg.observed
- wrg.tactic.code_review

[[ currentRule.title && currentRule.title.length > 60 ? currentRule.title.slice(0, 60) + '…' : currentRule.title ]]

f2a03bcd-ee82-468c-ba25-4106679d67a9

Propose edits, view related rules, and engage with the community through comments.

Executable code, not a declarative rule

This submission is arbitrary [[ currentRule.format ? currentRule.format.toUpperCase() : '' ]] code, not a declarative detection rule — review it as you would any script before running it, and note that it cannot be safely aggregated or auto-executed alongside YARA/Sigma/Suricata/CRS rules.

Rejected pattern present

[[ reason ]]

What does this mean?

Format & documentation suggestions

Not a security issue — this rule is valid and was accepted as-is. These are optional fields/conventions its own format recommends, mainly affecting its quality score and how well it documents itself for other reviewers.

[[ reason ]]

What do these messages mean?

[[ currentRule.title && currentRule.title.length > 60 ? currentRule.title.slice(0, 60) + '…' : currentRule.title ]]

[[ currentRule.title ]]

Community Discussion