GitHub Trend Radar v0.6.1

GitHub Trend Radar v0.6.1

This patch release fixes one real GitHub Trending edge case without adding new product scope.

Fixed

• Accept legitimate 0 stars today or 0 stars this week values as valid ranking data.
• Continue to fail closed when the period-star field is missing, malformed, or negative.
• Allow longer Top N requests on smaller programming-language lists to complete when a lower-ranked repository has zero period growth.

Validation

• 38 automated tests on Python 3.9 and 3.11 locally.
• Ruff lint, bytecode compilation, and strict v0.6.1 release validation.
• Real C# Top 20 fetch and ranking: 15 repositories returned, including dotnet/sdk with zero period stars and objective heat 0.0.
• GitHub CI passed for both main and the v0.6.1 tag, including Linux, macOS, and Windows jobs.

See the English README, 中文 README, and changelog.

GitHub Trend Radar v0.6.0

GitHub Trend Radar v0.6.0

This release turns profile portability, diagnostics, data retention, report quality, and public contribution support into complete, documented workflows.

Highlights

• Added confirmation-protected profile import with dry-run previews, idempotent merge, and explicit replacement mode.
• Added a separate feedback-history purge that preserves active preferences.
• Added read-only local and network diagnostics with doctor.py.
• Added preview-first retention cleanup for rebuildable snapshots and public metadata cache entries; profile and feedback files are always protected.
• Added conservative, auditable topic aliases for labels such as ai-agents, devtools, and llms.
• Fixed programming-language URL handling for names such as C# and Visual Basic, while preserving fail-closed validation for the requested Top N.
• Renamed GitHub's combined Issue and pull-request count to open_issues_and_pull_requests.
• Added real historical snapshot, redacted profile, and bilingual report examples, plus report-quality and behavioral evaluation guides.
• Added macOS and Windows compatibility jobs, a weekly live Trending parser smoke test, and structured Issue/PR templates.
• Expanded both READMEs into complete tutorials for installation, everyday use, profile migration, diagnostics, retention, recovery, and troubleshooting.

Install

npx skills add yuzilan/github-trend-radar --skill github-trend-radar --global

Requires Python 3.9 or newer and no third-party Python packages.

Validation

• 37 automated tests on Python 3.9 and 3.11 locally
• Ruff lint and bytecode compilation
• Strict release metadata and tag validation
• GitHub CI passed on the release commit before tagging
• Live GitHub Trending diagnostics and a real C# leaderboard fetch
• End-to-end export, dry-run import, merge import, and history-purge verification

See the English README, 中文 README, and changelog.

GitHub Trend Radar v0.5.0

GitHub Trend Radar v0.5.0

This release makes the skill easier to install, safer to use with public repository content, faster on repeated runs, and more predictable as it learns preferences.

Highlights

• Rewritten English and Chinese READMEs with copyable global installation, first-use prompts, expected output, preference controls, privacy details, updating, removal, and troubleshooting.
• Corrected personalization scaling: one weak repository investigation no longer receives a full interest bonus, and Top 1 requests are no longer replaced by the exploration slot.
• Added user-controlled forget-repo, forget-topic, confirmation-protected reset-profile, and portable feedback/profile export.
• Added interleaved daily/weekly metadata enrichment, a 24-hour public-metadata cache, concurrent GitHub requests, retry handling, and cache/API audit counts.
• Added an explicit trust boundary: GitHub pages, READMEs, issues, releases, and API fields are untrusted evidence, never agent instructions. Investigating a repository does not install or run it.
• Added --programming-language while retaining --language as a compatible alias.

Install

npx skills add yuzilan/github-trend-radar --skill github-trend-radar --global

Requires Python 3.9 or newer and no third-party Python packages.

Validation

• 26 automated tests
• Ruff lint
• Release metadata and tag validation
• CI on Python 3.9, 3.11, and 3.13
• Clean installation from the public GitHub repository verified against the tagged source

See the English README, 中文 README, and changelog.

GitHub Trend Radar v0.4.0

GitHub Trend Radar v0.4.0

A transparent Codex Agent Skill for manually discovering GitHub daily and weekly
Top 10 repositories, explaining them in plain language, and learning only from
explicit, inspectable user feedback.

Highlights

• Separate objective popularity from personalized relevance.
• Explain selected repositories from primary project sources.
• Record repository and topic interests, exclusions, and restoration events.
• Validate GitHub Trending HTML and fail closed on incomplete data.
• Protect local preference state with locking, atomic writes, and backups.
• Use only the Python standard library at runtime.
• Include English and Simplified Chinese documentation.

Install

npx skills add yuzilan/github-trend-radar --skill github-trend-radar

Python 3.9 or newer is recommended.

Verification

• Fresh installation from GitHub verified with the Skills CLI.
• 17 automated tests passed from the isolated installed copy.
• CI passed on Python 3.9, 3.11, and 3.13.
• Ruff and strict release metadata checks passed.

The skill never schedules itself, stars repositories, or installs and executes
trending project code without a separate explicit request.