Zero-dependency TypeScript Canvas library for full-link trace and user-behavior waterfalls: zoomable time axis, causal tree, sibling folding, virtualized rows.

tracelane scaffolding CLI. `npx @tracelane/cli init` detects your test runner (WDIO / Playwright / Cypress) and wires the matching adapter (@tracelane/wdio or @tracelane/playwright) into your runner config in one command — the published path from `npm ins

Framework-agnostic rrweb recording engine for tracelane. Wraps a per-framework browser object behind a common BrowserExecutor interface.

WebdriverIO Service that records rrweb sessions and writes a self-contained HTML report on failed tests. The user-facing tracelane integration for WDIO.

Self-contained, offline HTML report builder for tracelane. Embeds the rrweb player and a gzipped event blob into a single .html file.

Tracelane CLI — tlane command-line tool

Shared rrweb-based capture substrate for tracelane and peek.

Reference verifier for Tracelane tamper-evident audit ledgers

Tracelane TypeScript SDK — auto-instrument AI agent frameworks

Playwright reporter + auto-fixture that writes a self-contained replayable HTML report on failed tests — or every test in `all` mode (rrweb + console + failed-network). No SaaS.

Agent run tracing — spans, tool calls, cost, and failure receipts for production agents

Repo constitution / engineering taste packs for AI coding agents — constraints that stick

Lightweight eval harness for agent outputs — assertions, rubrics, and pass/fail suites

Agent policy gate — allow/deny tool actions, redact secrets, enforce permission boundaries

Tool-call reliability for agents — retries, backoff, circuit breakers, and failure classification

Prompt versioning and feedback patches for AI copilots — learn from thumbs-down without full fine-tunes

Schema and semantic handoff validation between agent steps — stop silent drift