Scan Report
0 /100
agent-session-state
Per-channel session isolation, write-ahead logging (WAL), and compaction recovery for AI agents.
This is a purely documentation-based skill for session state management with no executable code, scripts, or potentially dangerous behavior.
Safe to install
This skill is safe to use. It contains only documentation describing file-based session state management patterns.
| Resource | Declared | Inferred | Status | Evidence |
|---|---|---|---|---|
| Filesystem | READ,WRITE | READ,WRITE | ✓ Aligned | SKILL.md describes reading/writing session files in memory/sessions/ and memory/… |
| Network | NONE | NONE | — | No network operations described |
| Shell | NONE | NONE | — | No shell execution - only documented bash commands as examples for humans |
| Environment | NONE | NONE | — | No environment variable access |
| Skill Invoke | NONE | NONE | — | No skill invocation capabilities |
| Clipboard | NONE | NONE | — | No clipboard access |
| Browser | NONE | NONE | — | No browser access |
| Database | NONE | NONE | — | No database access |
File Tree
1 files · 5.7 KB · 160 lines Markdown 1f · 160L
└─
SKILL.md
Markdown
Security Positives
✓ Single-file documentation skill with no executable code
✓ Clear, well-structured documentation of intended behavior
✓ No scripts, binaries, or dependencies
✓ No network requests or data exfiltration
✓ No credential access or sensitive file manipulation
✓ No obfuscation or suspicious patterns
✓ Filesystem operations limited to user-controlled memory/ directory
✓ No supply chain risks (no package dependencies)