Hao Wang
Community@wanghao9610 · ShenZhen, China
Leave the world a little different.
Agent Skills by Hao Wang
Showing 32 vetted skills indexed across 2 GitHub repositories.
wanghao9610stage-outl-planner
Converts a finalized paper story into a compilable LaTeX manuscript skeleton with page budgets.
wanghao9610stage-resp-writer
Drafts point-by-point reviewer responses with a claim-evidence ledger and tracked promises.
wanghao9610stage-sect-drafter
Drafts one manuscript section per run from outline briefs and fingerprinted evidence.
wanghao9610stage-refs-curator
Curates verified BibTeX entries, reading notes, and reference indexes from fetched bibliographic records.
wanghao9610stage-proj-adopt
Migrate existing LaTeX manuscripts into the STAGE project layout with confirmed move plans.
wanghao9610stage-evid-curator
Import, register, and audit read-only research evidence files against a manifest.
wanghao9610stage-tabs-builder
Generate booktabs LaTeX tables from fingerprinted evidence files with per-row source comments.
wanghao9610stage-figs-designer
Builds and audits manuscript figures from editable tikz, python, or drawio sources.
wanghao9610stage-clms-auditor
Audits every manuscript number against fingerprinted evidence files and updates the claim ledger.
wanghao9610stage-pstr-builder
Plans, renders, and validates conference posters from verified manuscript claims and figures.
wanghao9610stage-copy-editor
Polish LaTeX manuscript prose while preserving every number, citation key, and todo marker.
wanghao9610stage-peer-reviewer
Simulates a five-perspective peer review panel that scores manuscripts against anchored rubrics.
wanghao9610stage-stry-coach
Coaches researchers through dialogue to define a paper's pitch, contributions, claims, and target venue.
wanghao9610stage-cite-auditor
Audit citation keys, cited-work assertions, and bib hygiene against reading notes.
wanghao9610stage-flow-status
Reports read-only status of manuscript sections, claims, evidence, and build with one next action.
wanghao9610stage-subm-packer
Preflight, package, and freeze academic paper submissions with build and lint gates.
wanghao9610stage
Routes /stage requests to the correct STAGE academic-writing workflow skill.
wanghao9610star-idea-storm
Guide researchers through structured ideation to produce defensible research topics.
wanghao9610star-plan-decomposer
Decompose research strategies into executable sub-plans with dependency graphs.
wanghao9610star-proj-adopt
Integrates existing AI research projects into the STAR framework via symlinks and historical logs.
wanghao9610star-code-reviewer
Audits source code for research-convention and plan-implementation compliance with evidence-backed reporting.
wanghao9610star-code-architect
Clone reference implementations and generate codearc.md architecture specs.
wanghao9610star-code-release
Consolidate research code and documentation into a public-ready repository format.
wanghao9610star-env-builder
Build and verify Python runtime environments with conda or venv backends.
Frequently Asked Questions About Hao Wang
FAQPage SchemaWhat tasks can I accomplish with wanghao9610's STAR and STAGE skills?▼
STAR covers research ideation, plan decomposition, environment building, code review, experiment auditing, and bibliography verification via DBLP, Crossref, and Semantic Scholar. STAGE covers manuscript outlining, section drafting, figure and booktabs table building, citation and claims audits, simulated peer review, rebuttal writing, poster rendering, and submission packaging.
Who are these skills designed for?▼
AI and computer science researchers, graduate students, and academic authors preparing conference or journal papers. They suit anyone managing the full pipeline from research planning and experiment validation through evidence-traced LaTeX writing, reviewer responses, and camera-ready submission packaging.
How do the STAGE manuscript skills work in practice?▼
Each skill is invoked via a /stage-* command with an argument hint. A typical run flows from story coaching to outline planning, section drafting, figure and table building from fingerprinted mates/ evidence, then claims and citation audits, mock review, response drafting, and final submission packaging with a git freeze tag.
How do the skills guarantee numbers in the paper are trustworthy?▼
Evidence lives in a read-only mates/ store with checksums and pinned upstream commits. Tables and figures only draw numbers carrying % src: comments to fingerprinted entries; missing values become \todo markers. The claims auditor traces every manuscript number and issues matched, mismatched, or unsourced verdicts.
What prerequisites and dependencies do these skills require?▼
STAR environment building uses conda or venv backends for Python runtimes. STAGE requires a LaTeX manuscript tree with main.tex, section inputs, and a BibTeX file; adoption of existing drafts or Overleaf exports is supported read-only. Bibliography verification fetches records from DBLP, Crossref, and Semantic Scholar.