v0.2.0

What's Changed

• Add verbosity/format CLI flags, revamp logging init, migrate StderrReporter by @hsubra89 in #170
• Remove unnecessary #[allow(dead_code)] on Error enum by @hsubra89 in #172
• Remove write-only StateManager by @hsubra89 in #171
• Extract rlph-core crate with pure domain types and algorithms by @hsubra89 in #173
• Rename rlph → brrr by @hsubra89 in #174

Full Changelog: v0.1.9...v0.2.0

v0.1.9

What's Changed

• Change default Codex model to gpt-5.4 by @hsubra89 in #156
• Unify CRITICAL tasks as batch-of-1 by @hsubra89 in #157
• Remove dead run_fix one-shot codepath by @hsubra89 in #158
• Reuse single shared worktree across fix batches by @hsubra89 in #159
• Log runner config when review/fix starts by @hsubra89 in #160
• Improve GitHub comment formatting for review output by @hsubra89 in #161
• Fail hard on inline review submission errors by @hsubra89 in #162
• Restructure as mixed-language monorepo by @hsubra89 in #165

Full Changelog: v0.1.8...v0.1.9

v0.1.8

What's Changed

• Upgrade ureq v2→v3 and toml v0.8→v1 by @hsubra89 in #119
• Show context window usage % in streamed agent output by @hsubra89 in #110
• Update README with current capabilities by @hsubra89 in #120
• Diff position mapper module by @hsubra89 in #128
• Remove internal review-fix loop and dead code by @hsubra89 in #130
• Post findings as inline review comments by @hsubra89 in #131
• Reaction-based fix workflow by @hsubra89 in #133
• Remove inert review_fix config/prompt path after orchestrator auto-fix removal by @hsubra89 in #134
• Resolve old review threads on re-review by @hsubra89 in #136
• Fix dry-run for reaction-based findings by @hsubra89 in #137
• Move build loop under rlph build subcommand by @hsubra89 in #139
• fix_scheduler module with ScheduleAction enum and unit tests by @hsubra89 in #145
• Add context usage % to Codex runner streaming output by @hsubra89 in #141
• Fix context % display to only track root agent usage by @hsubra89 in #146
• Wire scheduler into fix loop for CRITICAL findings by @hsubra89 in #147
• Fix integration tests to use build subcommand by @hsubra89 in #148
• Session batching for WARNING/INFO findings via RunBatch by @hsubra89 in #149
• Make --runner/--agent-binary/--agent-model global across all subcommands by @hsubra89 in #153
• Add type-safe ID newtypes, Duration config, CI caching, module docs by @hsubra89 in #152
• Review prompts: require suggested fixes, actionable findings, and consolidation by @hsubra89 in #154

Full Changelog: v0.1.6...v0.1.8

v0.1.6

What's Changed

• Add worktree setup script support by @hsubra89 in #117
• Use PR base branch for review diffs & show branch info by @hsubra89 in #118

Full Changelog: v0.1.5...v0.1.6

v0.1.5

What's Changed

• Embed finding JSON in review comments by @hsubra89 in #105
• Comment parser, updater, and dry-run CLI for rlph fix by @hsubra89 in #106
• Increase default implement phase timeout to 30 minutes by @hsubra89 in #108
• Single fix agent end-to-end (#101) by @hsubra89 in #107
• Parallel fix execution with push conflict retry by @hsubra89 in #109
• Continuous polling loop and graceful shutdown by @hsubra89 in #111
• Always clean up worktrees even on failure by @hsubra89 in #113
• Fix pre-commit hook for worktrees by @hsubra89 in #114
• Dependency tracking for fix tasks by @hsubra89 in #112
• Extract generic Tarjan SCC into shared module by @hsubra89 in #115
• Deduplicate retry_with_backoff across source modules by @hsubra89 in #116

Full Changelog: v0.1.4...v0.1.5

v0.1.4

What's Changed

• Add category to ReviewFinding, rewrite style review as sub-agent coordinator by @hsubra89 in #94
• Generate GitHub review comment from structured findings by @hsubra89 in #97

Full Changelog: v0.1.3...v0.1.4

v0.1.3

What's Changed

• Add codex JSON schema integration tests and stabilize binary CLI tests by @hsubra89 in #89
• Stream internal agent messages during review by @hsubra89 in #90
• Stream Codex internal agent messages during review by @hsubra89 in #91

Full Changelog: v0.1.2...v0.1.3

v0.1.2

What's Changed

• Extend reporter to full pipeline by @hsubra89 in #65
• Aggregator structured JSON output (replace magic strings) by @hsubra89 in #64
• Review phase structured JSON output + findings rendering by @hsubra89 in #66
• Fix agent structured JSON output by @hsubra89 in #67
• Add integration tests for Claude and Codex CLI binaries by @hsubra89 in #69
• Add integration tests for opencode CLI binary by @hsubra89 in #70
• Malformed JSON recovery via session resume by @hsubra89 in #68
• Add OpenCode agent binary support by @hsubra89 in #72
• Accept GitHub PR URLs in review command by @hsubra89 in #74
• PRD: Add formatting and lint checks to CI workflow by @hsubra89 in #75
• Fix .ralph → .rlph task.toml path typo by @hsubra89 in #84
• Auto-select single eligible task, skip choose phase by @hsubra89 in #85
• Add comprehensive Claude stream-json event schema tests by @hsubra89 in #88

Full Changelog: v0.1.1...v0.1.2

v0.1.1

What's Changed

• Add CLI binary integration tests (assert_cmd) by @hsubra89 in #57
• Add structured review output and suppress process noise by @hsubra89 in #58

Full Changelog: v0.1.0...v0.1.1

v0.1.0

What's Changed

• feat: CLI skeleton + config + trait definitions (#2) by @hsubra89 in #11
• feat: GitHub task source + dependency parsing (#4) by @hsubra89 in #16
• feat: Process module + bare Claude runner (#3) by @hsubra89 in #18
• feat: bug: missing default config file should fall back to defaults (#17) by @hsubra89 in #23
• feat: Worktree management + state tracking + prompt engine (#5) by @hsubra89 in #24
• fix: add flock-based file locking to StateManager by @hsubra89 in #25
• feat: Orchestrator + GitHub submission — first end-to-end loop (#6) by @hsubra89 in #27
• feat: bug: GitHub workflow labels should be matched case-insensitively (#22) by @hsubra89 in #28
• feat: bug: validate CLI overrides in effective config (#19) by @hsubra89 in #29
• fix: cycle tasks should still respect external blockers (#21) by @hsubra89 in #26
• Runtime hardening: process lifecycle, GitHub state flow, and base-branch support by @hsubra89 in #33
• doc: Add README.md file by @hsubra89 in #35
• fix: wire dependency graph into orchestrator to filter blocked tasks by @hsubra89 in #36
• feat: resume agent session on timeout by @hsubra89 in #38
• OpenAI Codex CLI runner by @hsubra89 in #40
• Add multi-platform GitHub Release workflow (tag + workflow_dispatch) by @hsubra89 in #42
• Rename BareClaudeRunner to ClaudeRunner and update runner flag values by @hsubra89 in #43
• Follow-up to #42: harden release dispatch and stdin error handling by @hsubra89 in #44
• Loop modes: continuous, max-iterations by @hsubra89 in #46
• Linear task source by @hsubra89 in #48
• Fix init CLI globals and config resolution for Linear setup by @hsubra89 in #51
• Make worktree creation fail on fetch failure with retries by @hsubra89 in #53
• Add rlph prd subcommand for interactive PRD creation by @hsubra89 in #52
• Replace single-agent review with parallel multi-phase review by @hsubra89 in #55
• Add rlph review command for direct PR review pipeline by @hsubra89 in #56

New Contributors

• @hsubra89 made their first contribution in #11

Full Changelog: https://github.com/hsubra89/rlph/commits/v0.1.0