Hidekazu Nagamura
Community@hideh1231
Ph.D student, Doshisha University
Agent Skills by Hidekazu Nagamura
Showing 9 vetted skills indexed across 1 GitHub repositories.
hideh1231develop-paper-outline
Builds and validates paper outlines as argument databases with a deterministic checker.
hideh1231read-paper
Guides closed-book reconstruction of a paper's argument and validates it with a deterministic checker.
hideh1231delegate-implementation
Guides Loopbench tasks from contract drafting through human-gated acceptance and implementation.
hideh1231work-task
Guides Loopbench tasks from contract acceptance through gated verification to human-approved completion.
hideh1231scan-decisions
Extract and classify project decisions from Loopbench records into an evidence-backed report.
hideh1231comprehend-contract
Guides human comprehension checks before Loopbench contract acceptance and human gate decisions.
hideh1231loopbench-retrospective
Inspect Loopbench task histories and propose reversible workflow improvements from evidence.
hideh1231challenge-decision
Reviews Loopbench contracts, human gates, and commit-sized work units with adversarial decision memos.
hideh1231setup-loopbench
Installs, repairs, and audits the standalone Loopbench CLI for the loopbench-skills plugin.
Frequently Asked Questions About Hidekazu Nagamura
FAQPage SchemaWhat tasks can I accomplish with hideh1231's skills?▼
You can execute non-trivial coding, debugging, and research tasks through Loopbench contracts with human gates, audit project decisions, challenge decisions with adversarial review memos, build paper outlines as argument databases, and reconstruct a paper's argument closed-book before viewing any generated summary.
Who are these skills designed for?▼
They target researchers, Ph.D students, and engineers running Loopbench-managed projects who need structured decision-making, external verification, and human approval gates. Academic users benefit from the paper outlining and critical reading skills built around deterministic argument checking.
What are the installation and runtime requirements?▼
Most skills require Python 3.11+, the loopbench CLI, and a project initialized with `loopbench init`. The setup-loopbench skill installs or repairs the CLI using uv. Contract and delegation skills additionally need a host capable of asking the user structured questions.
Are these skills open source and free to use?▼
Yes. All nine skills are released under the MIT license, permitting free use, modification, and redistribution. Loopbench itself is optional for the paper outline and reading skills, whose deterministic checker runs standalone.
When should I use challenge-decision versus work-task?▼
Use work-task to plan, execute, and finish a Loopbench task through external verification. Use challenge-decision when a commit-sized unit is ready, stakes are high, results are surprising, or you want an adversarial rubber-duck review producing a bounded memo without making the decision for you.