Three Ways to Record AI Behavior
Each method offers different confidence, coverage, and setup complexity. Use them independently or combine for maximum assurance.
- Records every tool call with arguments and results
- Zero changes to AI agent behavior
- Works with any MCP-compatible tool or AI gateway
- Cryptographically signed reports
- Bot account identification by platform
- Cryptographically signed webhook payloads
- Real-time event delivery
- Covers commits, PRs, reviews, deployments
- No additional setup — runs automatically
- Detects commit trailers and PR body markers
- Branch and workflow pattern matching
- Multiple signals increase confidence
What Gets Recorded
Research Foundation
LLM Exposure Monitoring: Platform Openness & Recording Depth
Systematic analysis of AI platform recording capabilities. Evidence basis for the attestation source model and confidence tier classification.
DOI: 10.5281/zenodo.19112060Our infrastructure boundary
OpenExecution is passive recording infrastructure. We receive events that platforms and tools push to our webhook endpoints — we do not instrument, modify, or extract data from tools that don't. Some AI agents (Claude Code, Cursor) maintain rich internal telemetry but provide no outbound event channel for third-party monitors. We are aware of this data and have the technical capability to build extraction tools, but we deliberately choose not to: infrastructure records what is pushed to it; extraction is outside our scope. We publish webhook API specifications and SDKs so that platform vendors, the open-source community, or users themselves can build bridges when they choose to.
OpenExecution records only. It never creates, modifies, or deletes resources in your connected platforms.