·
Last changedVersion:
This version was not hardened: a scanner flagged hostile content that hardening could not delete, so no model rewrote the skill and it is published without hardened content. Classified findings before hardening: 6. Classified findings closed: 0. Classified findings remaining: 6. Unclassified scanner findings remaining: 0. Review findings closed: 0. Review findings remaining: 15.
3
1
2
Expand | File | Group | Subtype | Severity | Status |
|---|---|---|---|---|---|
SKILL.md | Privilege Escalation & Host / Cloud Compromise | Privilege escalation | High | Open | |
SKILL.md | Memory & Context Poisoning | — | High | Open | |
SKILL.md | Supply Chain & Install-Time Trust | Dependency or plugin compromise | High | Open | |
SKILL.md | Sensitive Data & Secret Exposure | Data exfiltration | Medium | Open | |
SKILL.md | Tool & MCP Manipulation | MCP configuration abuse | Low | Open | |
SKILL.md | Prompt Injection | Third-party content as instructions | Low | Open |
6 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/google/skills/cloud/gemini-agents-api:latest
Apache-2.0
HARDENING.md