Getting it into your agent
It runs from inside its repository, so the clone comes first — what it calls does not travel with the file alone.
git clone --depth 1 https://github.com/yha9806/academic-writing-toolkitnpx agentmods add skills/yha9806/academic-writing-toolkit/thesis-controlWrote this? Show the measurements
A badge with what this costs and how it scanned, read live from this page, so it follows the numbers instead of freezing them. Markdown for a README, HTML for a documentation site or a project page.
[](https://agentmods.dev/skills/yha9806/academic-writing-toolkit/thesis-control)<a href="https://agentmods.dev/skills/yha9806/academic-writing-toolkit/thesis-control"><img src="https://agentmods.dev/badge/skills/yha9806/academic-writing-toolkit/thesis-control/github.svg" alt="Measured on agentmods" height="20"></a>Or the 80×15 button, for a site that already has a row of RSS and ATOM ones. Only the verdict fits; the numbers stay here.
<a href="https://agentmods.dev/skills/yha9806/academic-writing-toolkit/thesis-control"><img src="https://agentmods.dev/badge/skills/yha9806/academic-writing-toolkit/thesis-control.svg" alt="Reviewed on agentmods" width="80" height="20"></a>- NVIDIA SkillSpector pass
What it costs to keep this loaded
Counted locally with the o200k_base tokenizer, which is exact for GPT models; Claude uses its own tokenizer and its counts differ. Treat this as one consistent yardstick across the catalogue rather than a bill. Prices are per million input tokens.
| Model | Per session | Once invoked |
|---|---|---|
| Fable 5.1 | $0.00057 | $0.04110 |
| Opus 5 | $0.00028 | $0.02055 |
| Sonnet 5 | $0.00011 | $0.00822 |
| Haiku 4.5 | $0.00006 | $0.00411 |
Grade A, and why
thesis-control scanned grade A with 0 findings against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured 3d ago.
A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.
Nothing flagged
None of the 26 patterns this scan looks for appear in this file: no shell pipes, no recursive deletes, no credential paths, no hidden text, no instruction-override or anti-refusal phrasing, no agent-config snooping. That is not a guarantee, it is the absence of the things that are checkable.
How it starts
The opening of the file, as written. The whole thing — 451 lines — stays where its author put it; the contents beside it link to each section on GitHub.
/thesis-control - Thesis Drift Control
Purpose
Prevent AI-assisted writing from becoming fluent but distorted. Use this before and after substantive thesis or manuscript edits when the risk is not spelling or style, but loss of author control: project-level reframing, deletion of the real-world task or intended use, a primary domain becoming a secondary example, a changed research object or question, auxiliary analyses becoming primary, widened claims, blurred section purpose, missing caveats, unsynchronised adjacent paragraphs, or local edits that weaken the paper spine.
Trigger Words
This skill activates on: thesis control, drift audit, edit contract, spine card, claim drift, author control, loss of control, scope creep, rewrite risk, /thesis-control.
Core Rule
Do not edit thesis prose until the project intent, current manuscript contract, global thesis audit, section spine, and intended local change form one explicit and traceable contract chain.
The contract must answer:
This edit is allowed to change [specific local issue] in [specific unit], while preserving [spine sentence], [scope boundary], [core claims], and [do-not-change items].
If this sentence cannot be written, stop and diagnose the section instead of rewriting it.
The control hierarchy is:
Author-approved Project Intent
→ Author-approved Manuscript Contract
→ Passed Global Thesis Audit
→ Section Spine Card
→ Edit Contract
→ Post-edit Drift Audit
A lower layer cannot amend a higher one. If the title, abstract, primary domain, research object, research question, contribution scope, or manuscript structure no longer matches the approved intent, stop. Revise or roll back the manuscript, or create a new explicitly approved intent version that preserves the earlier row as history.
Control Files
Choose one control profile and name it as canonical.
For a lightweight single-manuscript workflow, read
references/author_control_lightweight.md and use:
What ships with it
6 files beside SKILL.md in the same directory: the scripts, references and assets a skill reads on demand. Not counted in the per-session cost; read them before you install if any of them is executable.
- scripts/check_thesis_control.py 39 KB runs code
- scripts/project_intent_control.py 23 KB runs code
- scripts/scaffold_thesis_control.py 28 KB runs code
- scripts/thesis_control_io.py 7.3 KB runs code
- scripts/upgrade_thesis_control_project_intent.py 11 KB runs code
- scripts/upgrade_thesis_control_revision_tracking.py 14 KB runs code
What this file has done since we first saw it
Hashed on every crawl. A supply-chain change to an agent config is a question of when, not whether, so the history is kept rather than the latest state alone.
- 3d ago Changed 4167e78185b5
- 5d ago Changed · +77 lines · +10 tokens per session 0ee930614eef
- 10d ago First seen · 374 lines · 47 tokens per session scan A 5e460c4488fb
thesis-control is a skill published in the GitHub repository yha9806/academic-writing-toolkit (36 stars, last pushed 2d ago), licensed MIT. It adds 57 tokens to every session and 4,110 once invoked, about $0.0003 per session on Opus 5. A static security scan graded it A with 0 findings. No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-30.
Other skills, from other repositories
figure-table-checker
Check, improve, and ensure compliance of figures and tables in academic papers. Use this skill whenever the user needs to verify figure or table formatting, choose error bar types, check colorblind accessibility, ensure APA or journal-specific compliance, create CONSORT or PRISMA flow diagrams, or fix common figure…
statistical-reporting
Report statistical results accurately and completely for any academic discipline. Use this skill whenever the user needs to write up statistics in APA 7th or other styles, report effect sizes, describe power analyses, interpret Bayesian factors, check reporting against CONSORT/STROBE/PRISMA/ARRIVE guidelines, or…
reference-verifier
Verify the accuracy and integrity of academic references. Use this skill whenever the user needs to check whether citations are accurate, find DOIs, detect retracted papers, identify AI-hallucinated citations, find orphan or phantom references, verify author names or publication details, or run a pre-submission…
abstract-writer
Write, evaluate, rewrite, or improve academic abstracts for research papers, dissertations, conference submissions, or journal manuscripts. Use this skill whenever the user needs an abstract written or improved, asks about abstract structure, wants to know what to include in an abstract, or needs to understand the…
grant-proposal
Write, structure, and improve grant proposals for any funding body — including NIH, NSF, EU Horizon, Wellcome Trust, ERC, and UKRI. Use this skill whenever the user needs to write a grant application, improve a funding proposal, understand grant review criteria, draft a specific aims page, write a budget…
lab-report
Write, structure, and improve lab reports for any scientific discipline. Use this skill whenever the user needs to write a lab report, understand what goes in each section, improve their methods or results section, format figures and tables, report statistics correctly, or understand the conventions of scientific…