Academic Writing Toolkit skill
A skill set for academic writing: source notes, citation audits, claim checks and local deterministic validation.
- 54 ★ on the repository
- Last commit
- MIT
Minimal access
Why this level
- The installer downloads Python dependencies and creates a private environment on your machine.
- Checks help, but decisions on claims stay with the author.
Install
- Works in
In your terminal, with SkillFoxx CLI
npx skillfoxx add skills/academic-writing-toolkitDetects the agents on your machine, shows the access level and pins the version.
Other ways to install
Access and checks
How we reviewMinimal accessThe tool runs nothing and makes no network calls.
- The installer downloads Python dependencies and creates a private environment on your machine.
- Checks help, but decisions on claims stay with the author.
This is third-party code. Review the repository files before installing.
What it does
Academic Writing Toolkit (AWT) is a local, evidence-controlled system for academic work. Nine skills: `read`, `note`, `map`, `integrate`, `readers`, `review`, `audit`, `verify-refs` and `export`. It keeps one notes file per source with page anchors and an evidence-status field, a citation-fidelity audit, a claim ledger bound to sources, a contribution-positioning audit and a prose fingerprint against the project's literature. Every check exits non-zero on an empty target and reports what it read. Judgment stays with the author: a machine reading is labeled as the machine's, and an approval an agent types into an editable file counts as a record, not a gate. Tools run locally with no model key, and online reference metadata checks switch on only with the `--online` flag.
Who it is for. Graduate students and researchers writing papers and theses with agents who want to verify that a citation says no more than the source.
Good fit when
- Before submitting a paper you check that each claim stays within what its source says.
- Sources are read and each gets notes with page references.
- The contribution is positioned against what is already published in the field.
- The reference list is verified against metadata before submission.
Not a fit when
- You want an agent to write the text for the author, since the set leaves judgments to the human.
- You need a cloud manuscript storage service, which the project does not include.
- Your environment is neither Codex nor a compatible local agent with shared skill folders.
Example request
Read this source, create a notes file with page anchors and mark whether the paper was read in full or only the abstract.Limitations
The `main` branch now ships the skill catalog, audits and workspace scaffold, while earlier distributions, including the Workbench, the Codex plugin and the ChatGPT App, are retired and available by tag. The latest tag `v0.6.0-rc.2` is marked a pre-release. It needs Python 3.9 or newer and Node.js 22.12 or 24 and newer. Installing into Codex runs from a repository clone with `scripts/install-codex-skills.py --install-deps`. Enforcement of limits sits with the agent environment rather than the project.
How to disable. Delete the installed skills from `~/.agents/skills` and the private Python environment the installer created. Your manuscript workspaces live separately.
FAQ
Which skills are included?
Nine: `read`, `note`, `map`, `integrate`, `readers`, `review`, `audit`, `verify-refs` and `export`.
Is a model key needed?
No. Deterministic tools run locally, and online reference metadata checks switch on with the `--online` flag.
What does the citation audit check?
It reads the notes files and checks that a cited claim stays within what the source says, for example that it does not rest on an abstract as if it were the full text.
Why does an empty target give an error?
Every check exits non-zero on an empty target and reports what it read. That avoids a false "nothing found".
How do I install into Codex?
Clone the repository and run `python scripts/install-codex-skills.py --install-deps` to place skills in `~/.agents/skills`.
Discussion
No questions or reviews yet.
Related
SkillsA plugin of two skills for a research paper: planning, drafting and revising sections, plus a four-pass manuscript review
A skill set for academic work: paper review, pre-analysis plan check, grant proposal review, code audit and slide decks.
Eight skills for finished papers and theses in LaTeX and Typst: compile, references, language polish, cover letters and a reviewer-style audit
A Claude Code skill for research papers: a five-stage pipeline, section rules, a style audit and an independent red-team review.
Eight skills draft a CS, AI or ML paper section by section and halt any section that lacks supporting evidence.
Claude Code skills for the academic pipeline: research, paper drafting, peer review and citation checks
Comparison with similar tools
Cards that share the main search query, and how they differ.
| Tool | Official | Repository stars | Main difference |
|---|---|---|---|
| Academic Writing Toolkit (this card) | no | 54 | |
| Academic Research Skills | no | 56,682 | difference: `academic-research-skills` supports academic research, while AWT audits citation fidelity and contribution positioning locally |
| Academic Skills for Claude Code & Codex | no | 5,921 | difference: that set is general purpose, while AWT is built around source notes and a claim ledger |
README in short
The README explains why the set exists, how it fits into paper writing, installation for Codex and from source, the skill list and the distribution history.