uditgoenka/autoresearch v1.8.2

v1.8.2 — Stability fixes and documentation consistency

on GitHub

Stability & Documentation Patch

10 bugs fixed from a 50-iteration /autoresearch:debug audit. No new features — purely stability and cross-reference consistency.

Bug Fixes

• learn.md — Added missing Argument Parsing section (9 flags were silently ignored)
• SKILL.md — Fixed --budget flag semantic mismatch (was "LLM cost", now correctly "max findings" matching predict-workflow.md)
• results-logging.md — Completed status enum (added keep (reworked), no-op, hook-blocked)
• debug.md — Added --technique flag to argument-hint and parsing
• fix.md — Added --skip-lint flag to argument-hint and parsing

Documentation Consistency

• getting-started.md — Fixed command count "7" → "9"
• README.md — Added learn.md, learn-workflow.md, autoresearch-learn.md to repo structure tree
• COMPARISON.md — Added learn to architecture tree, commands table, features list, chaining section (9 commands)
• CONTRIBUTING.md — Added COMPARISON.md and guide/scenario/ to repo structure, file table, PR guidelines
• Guide badges — Fixed stale 1.7.0 → 1.8.2 in debug and security guides
• Release scripts — Added guide/scenario/ and COMPARISON.md to doc review checklist

Full Changelog: v1.8.1...v1.8.2