Back to search
Top Contributors
73
commits
Security middleware for AI Agents. Intercepts shell commands before execution using a multi-layer pipeline: binary allowlist, regex patterns, deterministic intent coherence mapping, and LLM semantic check as last resort.
Stars
7
Forks
0
Watchers
7
Open Issues
0
Overall repository health assessment
No package.json found
This might not be a Node.js project
73
commits
chore: sanitize code comments and remove internal debug notes
e307f95View on GitHubfix: robust binary extraction in annotator for OpenHands/SWE-agent formats
eedfbc9View on GitHubfeat: add file upload with format validation and hints to forensics UI
82c7549View on GitHubfeat: improve forensics UX with random samples, visual dots, and robust normalization
b8fd84fView on GitHubfeat: add intelligent JSONL normalization to forensics UI
3af2766View on GitHubchore: add temporary debug logging and improve JSON parsing in forensics UI
adac3bdView on GitHubfix: restore emoji icons in dashboard template and fix encoding issue
993d481View on GitHubfix: correct forensics report data mapping in UI template
f5d41d3View on GitHubfeat: add forensics tab to UI and integrate with Guardian API
789a0faView on GitHub