Titan Gate – cryptographic receipts for AI-assisted code changes
Cryptographic chain-of-custody for AI code changes, SOC2 CC6/7/8 built in.

Replaces audit spreadsheet hell with tamper-evident GitHub-sourced evidence packs.
Engineering teams preparing for SOC 2, ISO 27001, or EU AI Act compliance audits
Vanta · Drata · Secureframe
EvidentTrail connects via GitHub App and captures PR approvals, branch protection changes, CI results, and AI-assisted commits as structured, tamper-evident evidence mapped to specific controls (SOC 2 CC8.1, ISO 27001 A.8.32, etc.).
A few things that might be interesting to this crowd:
- AI-assisted code detection: we identify co-authored-by trailers, bot authors, and PR labels to flag AI-generated code, then verify a human reviewed it before merge - Branch protection drift: hash-based snapshots detect when someone quietly reduces required reviewers or removes a required status check - Evidence packs: SHA-256 manifest + PDF/CSV/JSON export so the evidence is tamper-evident at export time, not just at capture time
First launch, no users yet. Happy to answer questions about the implementation or the compliance angle.
Cryptographic chain-of-custody for AI code changes, SOC2 CC6/7/8 built in.
Automated restore drills prove backups actually work when auditors come knocking.
SHA-256 hash chains and SMT proofs make AI agent logs actually verifiable.
Cryptographic hash chains make agent action logs legally defensible—LangSmith and Langfuse don't do this.
Cryptographic hash chains make AI agent logs legally defensible when LangSmith can't.
Curated list of 68 gh extensions with local browser-based audit.