OWASP Agent Memory Guard Model Card
这是可索引项目报告证据页：它保留 OWASP Agent Memory Guard Model Card 的源材料入口、机制线索和限制提醒；正文仍需 reader/editor 与 academic public-copy review 后才能当作最终结论引用。
OWASP Agent Memory Guard Model Card
| Field | Value |
|---|---|
| Repository | OWASP/www-project-agent-memory-guard |
| Category | Agent Memory Poisoning Defense and Guard Layer |
| Stars / forks snapshot | 17 / 10 |
| Language | Python |
| License | Apache-2.0 |
| Raw capture | raw-github/owasp_www-project-agent-memory-guard.md |
| Updated by | hourly public metadata update, 2026-05-30 19:17 +0800 |
1. Role in Self Evolve
OWASP Agent Memory Guard is a runtime protection layer for agent memory operations, targeting memory poisoning, leakage, and integrity tampering. It matters because self-evolving agents need repeatable harness control, measurable feedback loops, and reusable skill procedures before claiming stable improvement.
2. Working Principle
screen every memory read/write through detectors -> enforce declarative security policy -> emit forensics-ready events and snapshots -> block persistent memory poisoning before it propagates across sessions
3. Evidence Path
web-observed GitHub page showed 17 stars, 10 forks, 140 commits, Apache-2.0 license, and README claims as OWASP reference implementation for ASI06 memory poisoning. Shell GitHub API access remained blocked by DNS and local gh auth was invalid, so this card treats the snapshot as web-observed rather than API-verified.
4. Teaching Use
Use this card to explain Agent Memory Poisoning Defense and Guard Layer: it shows how harness/runtime/benchmark layers convert agent behavior into reproducible and auditable engineering workflows.
5. Limits
The repository was not cloned in this iteration; no benchmark run, plugin install, workflow execution, or agent loop experiment was executed. Counts and claims are visible public-page/search signals unless independently revalidated later.