Mapping your scanner's rules to AVE IDs — share your crosswalk

[chaksaray]

AVE assigns stable behavioral classification IDs to agentic AI component findings.
If you maintain a scanner that detects prompt injection, tool poisoning, exfiltration,
or similar agentic attack classes, mapping your rule IDs to AVE IDs lets security
teams deduplicate findings across tools using a shared vocabulary.

We have published crosswalks for NVIDIA SkillSpector and ClawScan already:

• crosswalks/skillspector-to-ave.json
• crosswalks/clawscan-to-ave.json

If you maintain a scanner and want help building a crosswalk for your rules,
open an issue or comment here with your rule categories. We will help with the mapping.

The implementer guide covers how to add ave_id to your finding output:
docs/specs/ave-implementer-guide.md