·
Last changedVersion:
This version did not pass the hardening gate. Hardening applied 44 targeted changes, deleted the lines high-severity poisoning findings flagged in 2 files before any model saw the skill, and preserved all other text. Classified findings before hardening: 31. Classified findings closed: 3. Classified findings remaining: 28. Unclassified scanner findings remaining: 0. Review findings closed: 21. Review findings remaining: 0.
2
8
21
Expand | File | Group | Subtype | Severity | Status |
|---|---|---|---|---|---|
references/code-blocks.md | Prompt Injection | Direct prompt injection | High | Closed | |
references/tables.md | Prompt Injection | — | Medium | Open | |
references/tables-to-html.md | Prompt Injection | — | Medium | Open | |
references/pandoc.md | Privilege Escalation & Host / Cloud Compromise | Privilege escalation | Medium | Open | |
references/code-blocks.md | Prompt Injection | — | Medium | Closed | |
references/hugo.md | Privilege Escalation & Host / Cloud Compromise | Privilege escalation | Medium | Open | |
references/basic-markdown.md | Obfuscation & Evasion | Trigger-based activation | Medium | Open | |
references/jekyll.md | Privilege Escalation & Host / Cloud Compromise | Privilege escalation | Medium | Open |
31 findings
No skill file found.
Bring Chainguard's hardening pipeline to your first-party skills — custom rulesets, automated remediation, full audit trail.
skills.cgr.dev/public/github.com/github/skills/markdown-to-html:latest
MIT
HARDENING.md