Skip to content

LLM05: refresh reference links for 2026#4

Merged
rocklambros merged 1 commit intoGenAI-Security-Project:mainfrom
RicoKomenda:upgrade/LLM05-refresh-references
May 2, 2026
Merged

LLM05: refresh reference links for 2026#4
rocklambros merged 1 commit intoGenAI-Security-Project:mainfrom
RicoKomenda:upgrade/LLM05-refresh-references

Conversation

@RicoKomenda
Copy link
Copy Markdown
Collaborator

@RicoKomenda RicoKomenda commented Apr 23, 2026

Summary

  • Replace four outdated 2022-2023 references with current CVEs, hands-on labs, and authoritative standards
  • Fix heading year (:2025 -> :2026) and section name (Common Examples of Vulnerability -> Common Examples of Risk per _template.md)
  • Fix OWASP ASVS publisher label typo (AASVS -> ASVS)

Replaced references:

  • CVE-2019-20634 (2019) -> EchoLeak CVE-2025-32711 (CVSS 9.3, zero-click exfiltration in M365 Copilot)
  • Vague Snyk/LangChain post -> LangGrinch CVE-2025-68664 (CVSS 9.3, LangChain serialization injection)
  • 2023 ChatGPT plugin post + 2023 Markdown exfil post -> PortSwigger hands-on LLM output handling lab + GitHub Copilot RCE via prompt injection (Embrace The Red, 2025)
  • AI-hallucinates-packages post (off-topic for output handling) -> dropped

Added references:

  • OWASP AISVS C7: Model Behavior, Output Control and Safety Assurance
  • CWE-116: Improper Encoding or Escaping of Output

Replace four outdated 2022-2023 references with current material:
- EchoLeak (CVE-2025-32711, CVSS 9.3): zero-click exfiltration in M365 Copilot
- LangGrinch (CVE-2025-68664, CVSS 9.3): LangChain serialization injection
- PortSwigger hands-on lab: exploiting insecure output handling in LLMs
- GitHub Copilot RCE via prompt injection (Embrace The Red, 2025)

Add new references:
- OWASP AISVS C7: Model Behavior, Output Control and Safety Assurance
- CWE-116: Improper Encoding or Escaping of Output

Remove: CVE-2019-20634, vague Snyk/LangChain post, 2023 ChatGPT plugin
post, 2023 Markdown exfil post, AI-hallucinates-packages post.

Also fix: heading year (2025 -> 2026), section name (Common Examples of
Vulnerability -> Common Examples of Risk per template), ASVS publisher
label typo (AASVS -> ASVS).
@RicoKomenda RicoKomenda force-pushed the upgrade/LLM05-refresh-references branch from 7f31af8 to 5282406 Compare April 23, 2026 09:30
Copy link
Copy Markdown
Collaborator

@rocklambros rocklambros left a comment

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Approving — conformance check passes against documentation/style/{README,general,entries}.md:

  • Renames ### Common Examples of Vulnerability to ### Common Examples of Risk (aligns with template's required section name)
  • All updated references use the documented [Title](URL): **Publisher** format
  • US English, ATX headings, no level skips

Merging per project owner @rocklambros's instruction since you are the LLM05 entry lead — your authorship is the merge gate.

@rocklambros rocklambros merged commit 2f23a8c into GenAI-Security-Project:main May 2, 2026
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

2 participants