SkillsAdd

smithery/zhuchichi56

repo-release-curation

Curate and restructure a near-publication ML repository for clean release, reproducible runs, and minimal layout while preserving eval/data and paper assets.

Installation

$ npx skills add smithery/zhuchichi56 --skill repo-release-curation
View on skills.sh ↗ Website ↗

Similar popular skills

Related neighbors and high-traction skills in the same topics — useful to compare before installing.

dataset-curation Similar name
wshobson/agents

Prepare, format, and validate datasets for supervised fine-tuning and preference training. Use …

1.3K installs
test-suite-curation Similar name
petrkindlmann/qa-skills

>- Audit a whole regression suite and prune/restructure it with evidence: per-test coverage fin…

10 installs
newsletter-creation-curation Similar name
brianrwagner/ai-marketing-claude-code-skills

Industry-adaptive B2B newsletter creation with stage, role, and geography-aware workflows

411 installs
photo-content-recognition-curation-expert Similar name
curiositech/some_claude_skills

Expert in photo content recognition, intelligent curation, and quality filtering. Specializes i…

318 installs
More details

Agent compatibility

Declared targets from SKILL.md / docs. Unmarked agents are not listed — the skill may still install via the CLI.

Claude Code Not declared
Cursor Not declared
Codex Not declared
GitHub Copilot Not declared
Windsurf Not declared
Gemini CLI Not declared
Cline Not declared
OpenCode Not declared

Package contents

Files included with this skill beyond the listing page.

  • skill md SKILL.md 1,952 B
  • docs SUMMARY.md 186 B

History

  1. First recorded snapshot · 0 installs

SKILL.md

Repo Release Curation

Use this skill when asked to tidy, restructure, or package a near-publication repository for open-source release while preserving reproducibility.

Workflow

  1. Confirm preservation scope

- Always keep: acl26_instructDiff/ (paper source), eval/, data/. - Confirm any other keep/delete constraints before removing files.

  1. Minimize root layout

- Target root: acl26_instructDiff/, configs/, data/, eval/, src/, README.md, requirements.txt, setup.py, pyproject.toml, .gitignore. - Move all code into src/ and adjust imports accordingly.

  1. Consolidate tools

- Merge plotting/analysis into src/<tool_pkg>/analysis/. - If utility helpers exist, merge into package init.py or a single module to reduce file count.

  1. Packaging

- Add requirements.txt, setup.py, and pyproject.toml. - Provide console_scripts entry for CLI (instdiff). - Ensure pip install -e . works without extra steps.

  1. Docs (bilingual)

- Write a single README.md with Chinese + English sections. - Include pipeline overview, quick start, config usage, and actual experiment results. - Reference paper figures using local assets.

  1. Configs and outputs

- Keep configs under configs/<domain>/. - Default outputs go to runs/ and are ignored via .gitignore.

  1. AI trace removal

- Remove assistant-specific logs or meta files. - Remove temporary logs or internal notes not needed for users.

Checks

  • pip install -e . installs instdiff CLI.
  • README.md is the only root doc, bilingual, and includes results.
  • Root contains only required top-level folders.
  • eval/ and data/ are preserved untouched.