fix: inference.py reset

fix: inference.py

fix: satisfy openenv multi-mode validation

fix: add OpenEnv environment manifest

fix: add OpenEnv root reset and step aliases

docs: init proto

feat: add OpenEnv-compliant root inference runner

feat: add dynamic registry-driven agent tool runtime

test: comprehensive LLM extraction testing and report

fix: improve LLM extraction prompts and column name parsing

feat: identify working LLM models and improve debugging

fix: improve model router error handling and add debug logging

fix: handle provider/model format in router (strip prefix before calling provider)

feat: prioritize user-site extraction over external search, improve metric patterns for HN-style sites

feat: improve tool call display with emoji prefix in dashboard

fix: improve complex link extraction for Twitch-style markdown

feat: intelligent search-based navigation for trending/popular content

test: verify strict output contracts on 200-case matrix

test: validate 35 non-template existing domains across output types

test: run 6100-case template and non-template stress matrix

fix: enforce template-hint agentic flow and strict output schema

refactor: remove hardcoded strategy scrapers

feat: Improve LLM extraction prompts for better data quality

fix: Kill stale backend process and verify agentic scraping works

feat: Replace hardcoded scraping with truly agentic LLM-driven approach

fix: GitHub trending CSV output returns correct columns

fix: resolve bare-domain assets to direct URLs

feat: add validate.url and html.parse tool calls for comprehensive visibility

feat: add comprehensive plugin registry with 71 tools across 12 plugins

feat: add comprehensive tool call visibility in step accordion

feat: add tool call visibility in step accordion

test: add runtime agent verification for 56 site templates

fix: dedupe stream steps and exclude agent skills from plugins

docs: document site template architecture and APIs

test: add full site template matrix coverage

feat: add site template registry and agent integration

test: expand test coverage to 15 tests, all passing

docs: add rewards and CSV output test report

fix: proper rewards for all steps and clean CSV output

feat: enhanced step accordion UI with Lucide icons and test report

fix: intelligent navigation scraper with GitHub trending support

feat: implement intelligent agentic web scraper

fix: resolve scraper functionality and plugin issues

test: comprehensive ScrapeRL system tests - 100% pass rate

chore: move AI test report to docs/test

test: comprehensive AI provider testing with 90% success rate

docs: comprehensive README update with all new features and examples

test: add end-to-end scraper test script with successful validation

feat: add frontend WebSocket integration with episode progress tracking

feat: add WebSocket support for real-time scraper progress updates