dsh-thesis
Run the following command in DeepSeek Harness:
dsh plugin install Paloma966/dsh-thesis
Paste the following prompt into your AI chat to install this plugin:
Run dsh plugin install Paloma966/dsh-thesis in the DeepSeek Harness terminal to install the plugin; full source is available at https://github.com/Paloma966/dsh-thesis
About this plugin
The costliest failure in thesis writing is not word count—it is discovering the direction was wrong and rewriting entire chapters. dsh-thesis changes not the act of writing but the ten minutes before it: it interrogates your school template, advisor requirements, plagiarism threshold, and deadline into a falsifiable intent specification, then advances through nine stages (topic, proposal, literature, design, implementation, testing, writing, compliance, defense) with full audit trails and a hard user sign-off at every gate.
The plugin exposes 21 tools and 11 methodology skill packs: thesis_ingest parses docx, xlsx, pdf, and md materials into structured summaries, thesis_intake runs a one-question-at-a-time state machine with checkpoint resumption so no answer is ever lost, thesis_lit_search walks a real fallback chain from Semantic Scholar through Crossref so every citation carries a genuine DOI enforced by a zero-fabrication cache, thesis_build ships a built-in zero-dependency docx generator that works without pandoc, and the defense_code suite breaks an AI-written system into modules you can walk through point by point on defense day. Originality scanning is a local deterministic estimate with no paid-system dependency; reports are fully reproducible and challengeable.
Built for undergraduate and master's students who already have a school template, experimental data, and code, and need a finished draft that survives plagiarism checks and oral defense. Also suitable for advisors who want to audit stage compliance across students. It will not conjure a thesis from a blank page; it amplifies the work you are already doing and carries it to a defensible finish.
Use Cases
- Walk the full nine-stage pipeline from proposal to defense slides once templates and data are ready
- Break an AI-assisted system into verifiable modules and rehearse implementation trade-offs before oral exam
- Scan high-risk passages for local originality estimates, apply rewrite prescriptions, and re-verify reduction
Best For
- Undergraduate or master's students who need a complete draft that survives plagiarism checks and oral defense
- Advisors or TAs auditing stage compliance and citation authenticity across students
- Graduate researchers who have code and data but lack a structured writing workflow
Related Plugins
A method pack that makes AI coding agents plan against your real baseline, prove completion with fresh evidence, and reduce reworks and unsafe changes.
Turns the DeepSeek Harness session into a captain that builds a durable sub-agent team, splits goals into dependency-aware tasks, and coordinates work via direct messages and a live Web UI.
Gives coding agents design judgment, letting Claude Code, Cursor, and 70+ agents generate and iterate high-quality UI, presentations, and graphics right from the terminal.
Run the Pi ecosystem's plugins on DeepSeek Harness, unmodified, via a compatibility layer that implements Pi's public extension ABI on DSH's native services.