AgentStack
SKILL verified MIT Self-run

Academic Writing

skill-mblauberg-provenant-academic-writing · by mblauberg

Use for drafting, rewriting, condensing, or checking academic prose and LaTeX while preserving claims and citations. Not for general, engineering, or Australian legal prose; use the matching writing skill.

No reviews yet
0 installs
0 views
view→install

Install

$ agentstack add skill-mblauberg-provenant-academic-writing

✓ scanned · ✓ verified — works with Claude Code, Cursor, and more.

Security review

✓ Passed

No issues found. Passed automated security review. · v0.1.0 How review works →

  • Prompt-injection patterns
  • Secret / credential exfiltration
  • Dangerous shell & filesystem operations
  • Untrusted network calls
  • Known-malicious package signatures

What it can access

  • Network access No
  • Filesystem access No
  • Shell / process execution No
  • Environment & secrets No
  • Dynamic code execution No

From automated source analysis of v0.1.0. “Used” means the capability is present in the source — more access means more to trust, not that it’s unsafe.

Are you the author of Academic Writing? Claim this listing to set pricing, connect Stripe payouts, and keep 70% of every sale.
Sign up to claim

About

Academic writing

Write precise, concrete, restrained and defensible academic prose. This skill owns academic voice cleanup; do not also invoke a general writing skill unless the user asks. It does not cover source discovery, bibliography management, invented citations, integrity adjudication or assessment-permission advice. AI-use disclosure wording and placement are in scope; permission and policy are not.

The target venue, institution, supervisor-approved style, project glossary and local LaTeX/citation conventions take precedence over these defaults. Diagnosis is read-only. Edit only assigned files and do not restructure a work, change a bibliography or replace citations without explicit authority.

Workflow

  1. Identify the section and mode: draft, rewrite, condense, diagnose,

section-polish, match-voice, citation-safe or final-scrub.

  1. Lock claims, numbers, units, citations, quotations, LaTeX commands, equations,

labels, cross-references, result macros, paths and technical names.

  1. This skill is a specialization of the natural-writing hub: it owns

thesis register, LaTeX/citation-key preservation and the empirical-research overlays, and links to the hub for the shared doctrine. Load the natural-writing skill for its Australian English default and, for prose, [academic register](references/academic-style-au.md) plus the matching [workflow mode](references/editing-workflows.md); see that file's reference map for which remaining hub/overlay pair (anti-AI, condense, claim discipline, chapter, engineering voice or citation/LaTeX safety) a mode needs. A project adapter or keep-list takes precedence.

  1. Preserve meaning and evidence altitude. Separate observation, interpretation,

limitation and future work. Flag unsupported claims; never strengthen them.

  1. Follow the target language style (otherwise Australian English). Use concrete

verbs and purposeful sentence variety. Cut noun stacks, inflation, empty meta-discourse and instruction-leaking absence tombstones; apply punctuation defaults only where the project is silent. Keep honest negatives and defined technical terms. Do not edit to beat an AI detector.

  1. For a full chapter, paper or document rewrite or final polish, use the

risk-proportional independent review in editing-workflows.md. Load orchestrate for runtime routing; do not select providers or model IDs here. Preserve labels, macros and citation keys, then validate the whole work.

  1. Run one adversarial final scrub: name and repair only a real remaining tell.

For condensation, follow the reference stop rule, report words before/after and percentage cut, and verify every locked invariant. If another cut would lose a fact, caveat, scope condition, term or reference, stop.

For local files, use the checker as a review prompt, not proof:

python3 "${AGENTS_HOME:-$HOME/.agents}/skills/academic-writing/scripts/check_academic_style.py" path/to/file.tex

Return rewritten prose first and concise risk flags second. For diagnosis, return highest-risk findings first. Leave unsafe sentences unchanged and state why. Match surrounding project conventions when editing files.

Source & license

This open-source skill is cataloged on AgentStack and links to its original source — we do not rehost the code.

Install and usage instructions live in the source repository linked above.

Reviews

No reviews yet — be the first.

Versions

  • v0.1.0 Imported from the upstream source.