Cursor Plugins

AMD

by AMD
dda93f95-20cc-5d15-9af5-23888a2bfad9 | Store vb0c6f8030535
80/ 100
HIGH risk
Risk verdict
Review before use

Score-based assessment (high risk, 80/100). An AI review is available in the AI review tab.

Analysis record

Analysed
2 weeks ago 1 November 2026
Version
vb0c6f8030535
Artifact
SHA256 17f5a320…ca046d
Detail
Full list

Evidence ledger

Ranked by severity · findings with a source location link to the code viewer

355 findings

YARA Rule Matches

11 rules
SeverityRuleHitsFilesMetadata
LOWpostinstall network communication 35
walkthroughs/lemonade-router-builder.mdwalkthroughs/tracelens-analysis-orchestrator.mdeval/agent.py +32 more
-
LOWpostinstall system command 100
skills/tracelens-analysis-orchestrator/agents/triton-analyzer.mdskills/local-ai-app-integration/reference.mdstaging/rocm-doctor/evals/evals.json +97 more
-
LOWUsingShellInterpreterWhenExecutingOSCommands 1
skills/local-ai-app-integration/reference.md
-
LOWpostinstall obfuscation 68
skills/lemonade-router-builder/evals/evals.jsonskills/serving-llms-on-epyc/scripts/cpu_tune.pyskills/serving-llms-on-instinct/evals/hooks.py +65 more
-
LOWpostinstall crypto operations 2
skills/serving-llms-on-instinct/data/recipes_cache.jsonskills/local-ai-app-integration/reference.md
-
LOWpostinstall file manipulation 49
eval/test_evals.pywalkthroughs/lemonade-router-builder.mdskills/hyperloom-workload-optimizer/SKILL.md +46 more
-
LOWpostinstall registry modification 10
staging/apu-memory-tuner/SKILL.mdstaging/apu-memory-tuner/scripts/detect_platform.pystaging/apu-memory-tuner/scripts/apply_profile.py +7 more
-
LOWpostinstall file download 36
.github/workflows/evals.ymlskills/local-ai-use/reference.md.github/workflows/external-reference-check.yml +33 more
-
LOWcredential env files 28
skills/hyperloom-workload-optimizer/setup.mdskills/hyperloom-workload-optimizer/scripts/tests/test_preflight.pyeval/routing.py +25 more
-
LOWpostinstall persistence mechanism 14
skills/serving-llms-on-instinct/data/recipes_cache.jsonskills/hyperloom-workload-optimizer/SKILL.mdskills/local-ai-use/scripts/setup_local_ai.py +11 more
-
LOWpackage json suspicious scripts 1
skills/serving-llms-on-instinct/data/recipes_cache.json
-

Finding Categories

1
Network
400
IoC Indicators

YARA Rules Matched

11 rules(344 hits)
postinstall network communication postinstall system command UsingShellInterpreterWhenExecutingOSCommands postinstall obfuscation postinstall crypto operations postinstall file manipulation postinstall registry modification postinstall file download credential env files postinstall persistence mechanism package json suspicious scripts

AI Security Report

AI Security Review

Evidence context: threat category tool poisoning; evidence quality moderate.

Jev rated 8 findings against the supplied code: 1 likely real, 3 likely benign pattern matches. Likely real: SKILL-PROMPT-INJECTION-skills/local-ai-use/SKILL.md-6 (skills/local-ai-use/SKILL.md:6). The finding list was truncated; findings beyond it were not rated.

Key Reasons

  • SKILL-PROMPT-INJECTION-skills/local-ai-use/SKILL.md-6 (skills/local-ai-use/SKILL.md:6)
  • SKILL-EXCESSIVE-AGENCY-staging/rocm-doctor/SKILL.md-84 (staging/rocm-doctor/SKILL.md:84)
  • SKILL-PROMPT-INJECTION-skills/local-ai-use/SKILL.md-305 (skills/local-ai-use/SKILL.md:305)

False Positive Considerations

  • YARA--credential_env_files (.github/skillspector-allow.yml)
  • YARA--postinstall_obfuscation (.github/skillspector-allow.yml)
  • YARA--postinstall_persistence_mechanism (.github/lychee.toml)

Reviewed 2026-11-15 · version b0c6f8030535; recommended action: escalate; model confidence 52%.

Source Code Viewer

Source Code Not Available

Source code is not available for this version of the extension.

About This Extension

AMD's verified Agent Skills in one plugin: route image/audio through local AI on Ryzen AI, serve LLMs on AMD Instinct GPUs with vLLM, optimize inference throughput with Hyperloom, and analyze GPU kernel and PyTorch trace performance.

Frequently Asked Questions