Changelog
Unreleased¶
Added¶
-
add Pi backend and multi-turn conversation context by @shenxianpeng (96d8627)
-
enable Pi backend for PR fix mode and dogfooding by @shenxianpeng (8682f7b)
Changed¶
-
update deployed agent workflow to match template by @shenxianpeng (183083a)
-
use uses: ./
to call local actions in all workflows by @shenxianpeng (b84b965) -
install Node.js + Pi in agent action for pi backend support by @shenxianpeng (6cb6a8f)
-
simplify architecture diagram to 6-grouped-edge pipeline by @shenxianpeng (49476c5)
-
auto-detect local checkout for dogfooding install by @shenxianpeng (8e08353)
-
auto-sync CHANGELOG.md with git-cliff on every push by @shenxianpeng (ed2f5ae)
-
update README and docs for v1.3 — Pi, PR fix, inline review by @shenxianpeng (6b01e6b)
-
update CHANGELOG.md [skip ci] (b278b7a)
-
accurate 4-way comparison table with Copilot, CodeRabbit, PR-Agent by @shenxianpeng (3927e5b)
Fixed¶
-
stop auto-adding repokeeper-labeler on every labeled issue by @shenxianpeng (a376d35)
-
add actions/checkout before uses: ./
in all workflows by @shenxianpeng (1b391aa) -
handle empty PR_NUMBER env var gracefully by @shenxianpeng (4d13a4e)
-
robust Pi output parsing, clearer JSON instructions by @shenxianpeng (bf405da)
-
log Pi stderr, surface crash errors in issue comments by @shenxianpeng (52ffcd0)
-
set DEEPSEEK_API_KEY env for Pi deepseek provider by @shenxianpeng (a4b97b5)
-
pass --provider to Pi explicitly (defaults to google) by @shenxianpeng (8a3c90b)
-
warn about untracked files before Pi runs by @shenxianpeng (4097ebf)
-
replace git add -A with git add -u + explicit new file staging by @shenxianpeng (45c652c)
v1.2.0 - 2026-05-10¶
Added¶
-
inline PR review comments, describe, similar-issue detection, incremental re-review by @shenxianpeng (be9ad30)
-
Add codecov.yml with project/patch coverage thresholds and PR comments by @shenxianpeng (008ee90)
-
Add GitHub Discussions community badge and section to README and docs nav by @shenxianpeng (8368a32)
-
Add LLM API connectivity validation to repokeeper doctor by @shenxianpeng (d314b58)
-
Add repokeeper pricing command with staleness check and env override hints by @shenxianpeng (75c69d9)
-
Add JSON-structured logging via --log-format and RKP_LOG_FORMAT by @shenxianpeng (7dc4b93)
-
Add benchmarks documentation with cost/performance estimates per scenario by @shenxianpeng (5a89c78)
-
Add comprehensive tests for LLM client: Anthropic, streaming, JSON repair — coverage 56%→93% by @shenxianpeng (21bc107)
-
Add integration-level tests: agent pipeline, verification fix loop, branch collision — coverage 82%→84% by @shenxianpeng (df516cb)
-
add conversational PR fix mode by @shenxianpeng (e091ccc)
Changed¶
-
remove doctor action by @shenxianpeng (8ae43b6)
-
update docs for inline review, describe, similar-issue detection, incremental re-review by @shenxianpeng (e2c1860)
-
Revert "Add GitHub Discussions community badge and section to README and docs nav" by @shenxianpeng (1eaa64b)
-
make overview concise, put all issue details in inline comments by @shenxianpeng (1d8f3eb)
-
add Mermaid architecture diagram to README by @shenxianpeng (665785e)
-
redesign architecture diagram for vertical flow by @shenxianpeng (56a9d16)
-
change trigger from @repokeeper to /repokeeper by @shenxianpeng (f2e9174)
v1.1.0 - 2026-05-07¶
Added¶
- enhance implementation change handling and context expansion by @shenxianpeng (f86403f)
Changed¶
-
update AGENTS.md to refine conversational style and clarify git rules for parallel agents by @shenxianpeng (24eb74d)
-
auto-move v1 tag on release by @shenxianpeng (7940811)
Fixed¶
-
update deploy condition to include workflow_dispatch event by @shenxianpeng (b7797bb)
-
fix CI scan, dep count display, and add step logs to LLM diagnosis by @shenxianpeng (29e804b)
-
resolve mypy errors in repo_context.py, add type-check requirement to AGENTS.md by @shenxianpeng (e5838a7)
v1.0.0 - 2026-05-07¶
Added¶
-
add Auto-Labeler module (Module 5) with docs by @shenxianpeng (64bd0f0)
-
per-module AI model selection (labeler/radar/patrol/review) by @shenxianpeng (d3eb23d)
-
composite GitHub Actions — one action per module by @shenxianpeng (05a48a3)
Changed¶
-
chore(deps): bump taiki-e/install-action from 2.75.30 to 2.77.1 in the actions group by @dependabot[bot] in #20
-
add dogfood labeler workflow for auto-labeling issues and PRs by @shenxianpeng (1418375)
-
fetch repo labels once in batch mode instead of per-issue by @shenxianpeng (55b4fac)
-
update tool comparison — modernize Copilot/Cursor, add PR Agent by @shenxianpeng (ba4f919)
-
update CHANGELOG for v1.0.0 by @shenxianpeng (0a5b8ca)
Fixed¶
- resolve mypy any-return error in get_module_model by @shenxianpeng (ad2ddb5)
v0.10.0 - 2026-05-06¶
Added¶
-
two-step smart file selection and verification fix loop (ff602de)
-
add Code Review Agent (Module 5) (86fbd3e)
-
add review workflow for dogfooding (b837f44)
Changed¶
- update changelog for version 0.10.0 by @shenxianpeng (01a2a5f)
Fixed¶
- add mypy type ignore for check_review_skip_keywords (ea63fd4)
v0.9.0 - 2026-05-06¶
Added¶
- Add agent candidate handoff workflow by @shenxianpeng in #12
Changed¶
- update changelog for version 0.9.0 (2d0ee8a)
Fixed¶
- fix: 10 code quality and safety improvements by @shenxianpeng in #13
v0.8.0 - 2026-05-05¶
Added¶
- add cross-repo global search by @shenxianpeng (4e87f24)
Changed¶
-
add repokeeper badge and styling in README by @shenxianpeng (66a2eab)
-
update example project name in Community Radar documentation by @shenxianpeng (22df5fb)
-
add edge case tests for collect_repo_files and apply_and_push functions by @shenxianpeng (3788f58)
-
update changelog for version 0.8.0 with new features and documentation updates by @shenxianpeng (eab8db6)
v0.7.0 - 2026-05-05¶
Added¶
- auto-create issues from community hits with deduplication and branding by @shenxianpeng (eacd1d0)
Changed¶
-
update changelog for version 0.6.0 with new features and fixes by @shenxianpeng (fc1762c)
-
update changelog, readme, radar docs, and workflow for auto-create feature by @shenxianpeng (729678d)
v0.6.0 - 2026-05-05¶
Added¶
- make llm cost estimates configurable by @shenxianpeng (80cad48)
Changed¶
-
limit source distribution contents by @shenxianpeng (03457c0)
-
raise coverage gate by @shenxianpeng (2246a9b)
-
clarify workflow install paths by @shenxianpeng (cab1288)
Fixed¶
-
constrain llm file writes to repo by @shenxianpeng (873be6b)
-
update commit preprocessors and parsers for better categorization by @shenxianpeng (d0d67c2)
v0.5.0 - 2026-05-04¶
Added¶
- improve setup diagnostics by @shenxianpeng (dcf804d)
Changed¶
-
sync README workflow setup with quick-start guide by @shenxianpeng (f42f356)
-
document security model by @shenxianpeng (aaa7bf4)
-
add contribution templates by @shenxianpeng (1db86a8)
-
add doctor to quick start by @shenxianpeng (8e03e77)
-
add uv.lock by @shenxianpeng (10c200e)
Fixed¶
- include workflow templates in package by @shenxianpeng (00f5c88)
v0.4.0 - 2026-05-03¶
Added¶
- P4 LLM enhancements — Anthropic, streaming, token tracking, branch collision by @shenxianpeng (b6729a8)
Changed¶
-
rewrite adoption section with copy-workflow, CLI, and AI-prompt approaches by @shenxianpeng (dede402)
-
Replace release-drafter with git-cliff for draft releases by @shenxianpeng (18d2d4b)
-
Update cliff.toml section names to Keep a Changelog style by @shenxianpeng (1c95cf6)
-
comprehensive P1 test coverage — patrol 25%→87%, radar 30%→89%, total 67%→89% by @shenxianpeng (bed609b)
-
P2 architecture improvements — mypy, logging, split agent.py by @shenxianpeng (77f53d6)
-
prepare 0.4.0 release — changelog, README, all module docs by @shenxianpeng (4bca2b0)
Fixed¶
-
patrol report footer links to repokeeper repo, not the target repo by @shenxianpeng (25e8e5a)
-
implement all P0 features — discussion scanning, CI log fetching, CI auto-fix, multi-ecosystem dep checks by @shenxianpeng (f626a5b)
-
CI mypy errors — missing type stubs and Any-return in radar.py by @shenxianpeng (a190154)
v0.3.0 - 2026-05-03¶
Added¶
-
Add GitHub Actions workflow for RepoKeeper Agent by @shenxianpeng (7f0ad30)
-
Add initial configuration for RepoKeeper by @shenxianpeng (03d8577)
-
Add initial README with project title by @shenxianpeng (3dcb62e)
-
implement all four RepoKeeper modules with full documentation by @shenxianpeng (616084d)
-
prepare repokeeper for packaging and ci by @shenxianpeng (af2b14d)
-
add --summary flag and upload radar-summary artifact by @shenxianpeng (09536e5)
-
improve onboarding cli checks by @shenxianpeng (a75fdce)
-
verify agent changes before pr by @shenxianpeng (39c07cf)
-
inject CHANGELOG latest section into release draft body by @shenxianpeng (dd9f9e9)
Changed¶
-
Create agent.py by @shenxianpeng (38e0590)
-
Correct copyright holder name by @shenxianpeng (ab59ce1)
-
support python 3.10 through 3.14 by @shenxianpeng (c8e4443)
-
remove dead .github/repokeeper/agent.py wrapper (CLI handles this) by @shenxianpeng (859537f)
-
bump version to 0.2.0, add dependabot config by @shenxianpeng (5443524)
-
group all deps into single PRs by @shenxianpeng (8a80ee8)
-
chore(deps): bump the actions group with 5 updates by @dependabot[bot] in #7
-
add comprehensive tests achieving 95+% agent, 100% profile, 90% cli coverage by @shenxianpeng (400c9e3)
-
add AGENTS.md and fix all lint issues by @shenxianpeng (2d0d3c8)
-
redesign README and docs with AI-project feel, Copilot comparison, zero-config adoption by @shenxianpeng (9b95be5)
-
pipguard-style redesign with custom CSS, logo, hero section, terminal mockup by @shenxianpeng (121a8fd)
-
update documentation and assets for improved clarity and design by @shenxianpeng (9aac522)
-
align setup guide with current capabilities by @shenxianpeng (a04d941)
-
refresh mkdocs visual design by @shenxianpeng (5caeef9)
-
sync template workflow action versions with own repo by @shenxianpeng (5656526)
-
migrate to setuptools-scm for automatic versioning by @shenxianpeng (468cefb)
Fixed¶
-
block .github/workflows/ modifications to prevent push rejection by @shenxianpeng (92296c1)
-
handle 404 on get_repo with GITHUB_TOKEN fallback by @shenxianpeng (8cf4185)
-
write radar/patrol summary to GITHUB_STEP_SUMMARY by @shenxianpeng (aaedb8e)
-
auto-strip blocked paths instead of skipping whole issue by @shenxianpeng (73282cf)
-
add real-time logging and resilient LLM JSON parsing by @shenxianpeng (c3a4d2b)