Scan Report
5 /100
tiered-context-manager
多Agent协作的智能会话上下文管理系统。当需要管理AI agent的长会话压缩、多层记忆分层、跨Agent知识共享时激活。支持L1/L2/L3分层压缩、实时监控、统计分析。用于OpenClaw agent的上下文管理优化。
This is a legitimate multi-agent context management system with session compression, memory tiering, and cross-agent knowledge sharing capabilities. No malicious behavior detected.
Safe to install
This skill is safe to use. Consider documenting the filesystem:WRITE capability explicitly in SKILL.md for transparency.
Findings 1 items
| Severity | Finding | Location |
|---|---|---|
| Low | Missing filesystem write declaration Doc Mismatch | SKILL.md:1 |
| Resource | Declared | Inferred | Status | Evidence |
|---|---|---|---|---|
| Filesystem | NONE | WRITE | ✓ Aligned | SKILL.md: No declaration; tiered-engine.js reads/writes session files and create… |
| Network | NONE | NONE | — | No network requests found in codebase |
| Shell | NONE | NONE | — | No subprocess or shell execution found |
| Environment | NONE | NONE | — | No os.environ iteration or sensitive variable access |
File Tree
14 files · 114.8 KB · 3957 lines JavaScript 9f · 3330L
Python 2f · 383L
Markdown 2f · 235L
JSON 1f · 9L
├─
▾
references
│ └─
architecture.md
Markdown
├─
▾
scripts
│ ├─
compression_stats.js
JavaScript
│ ├─
cross_agent_context.js
JavaScript
│ ├─
index.js
JavaScript
│ ├─
l3_ai_compressor.js
JavaScript
│ ├─
memory_tiering.js
JavaScript
│ ├─
package.json
JSON
│ ├─
process_l3_tasks.py
Python
│ ├─
realtime_monitor.js
JavaScript
│ ├─
save_shared_state.py
Python
│ ├─
tiered_standalone_v2.js
JavaScript
│ ├─
tiered_standalone.js
JavaScript
│ └─
tiered-engine.js
JavaScript
└─
SKILL.md
Markdown
Dependencies 3 items
| Package | Version | Source | Known Vulns | Notes |
|---|---|---|---|---|
fs (Node.js built-in) | bundled | node | No | Standard filesystem module, no external dependencies |
path (Node.js built-in) | bundled | node | No | Standard path module, no external dependencies |
os (Node.js built-in) | bundled | node | No | Standard OS module, no external dependencies |
Security Positives
✓ No credential harvesting or API key theft detected
✓ No external network requests or C2 communication
✓ No base64 encoding, obfuscation, or eval() calls
✓ No curl|bash or wget|sh remote script execution
✓ Proper backup mechanism before file writes (safeWriteSession)
✓ Auto-rollback on write failures for data protection
✓ No access to sensitive paths (~/.ssh, ~/.aws, .env)
✓ No reverse shell, RCE, or data exfiltration patterns
✓ Clean codebase with no obfuscation techniques