Find and prevent secrets and prompt injection attacks in AI coding agent sessions
Deterministic, auditable, reproducible deception infrastructure for attacker interaction capture and analysis
Stage
Not given
Startup
Founded
Not given
2025
Based in
Not given
Not given
Pricing model
Open source
Open source
Pricing
Not given
Echidra OSS is free and open source under AGPLv3 license. Echidra X is a future enterprise layer currently in roadmap direction with early access available upon request.
Key features
Scan recorded AI agent sessions for known credentials
Guard against prompt injection attacks
Local console with audit trails at 127.0.0.1
Support for multiple AI coding agents including Claude Code, Cursor, and Codex CLI
Default policy with 21 rules across 19 action classes
Test policy against replayed attack scenarios
No model or network in the guard's hot path
Session capture and behavior analysis across SSH, HTTP, FTP, and Telnet
Automated MITRE ATT&CK mapping
Recurring security findings with recommended actions
Interactive fake shell and login prompts
Self-hosted, nothing leaves your infrastructure
Configurable personas to control attacker visibility
Behavior classification and actor labeling
Cross-session correlation for recurring issues
What makes it different
Deterministic policy-based enforcement rather than model-based decisions
Runs locally with no telemetry and no account required
Open source and free forever under Apache-2.0
Replayed 36 real attack scenarios with zero bypasses
Findings that group attacker sessions with recommended fixes
Local-first, self-hosted architecture with no data sent to external services by default
Deterministic, auditable, and reproducible classifications