nicsuzor
75 verified skills77 total stars
deep-research
Author high-quality deep-research prompts (Gemini / ChatGPT Pro / Perplexity Deep Research), then capture the resulting documents into the PKB — including figure extraction, agent-transcribed alt-text for load-bearing images, frontmatter, and wikilink wiring to the sourcing task.
testing2
cowork-sync
Mirror PKB tasks onto the Cowork native task list at claim time and sync completion back to PKB. Cowork-only; ships only in the cowork build of aops-core.
development2
analyst
Support academic research data analysis with technology-agnostic principles — research-data immutability, a versioned/tested/reproducible transformation layer, statistical methodology, and self-documenting research. Use this skill for any computational research project with an empirical data pipeline. The skill enforces academicOps best practices for reproducible, transparent research with a collaborative single-step workflow. Tech-specific how-to (dbt, Streamlit, Python plotting/stats) lives in the aops-extras package.
development2
python-viz
Python plotting and statistical-modelling libraries (matplotlib, seaborn, statsmodels) for the analyst presentation and statistical-methodology layers. Use when producing publication-quality figures or fitting statistical models in Python. Library-specific HOW for the tech-agnostic principles in the aops-tools analyst skill.
tools2
remember
Unified memory skill: immediate mode (/remember) persists knowledge via PKB MCP; maintenance mode (/sleep, GHA cron) runs periodic consolidation — transcript mining, knowledge synthesis, data quality, brain sync.
tools2
narrative-digest
Launder supervisor/worker task-log output into a Nic-facing narrative — what happened, where things are headed, and what (if anything) is genuinely his to decide. Never relays raw process detail (worker IDs, thread pointers, log paths) or verbatim task-log stream-of-consciousness.
testing2
dump
Emergency session bail — fast resume task + short handover, no commit/PR/reflection. For when you (or the user) need a clean context now. Use /end-session for canonical close.
data-ai2
daily
Daily note lifecycle — compose and maintain a factual daily note. Reports the state of the day; does not prioritise or recommend. SSoT for daily note structure.
data-ai2
extract
General extraction/ingestion skill that routes to specific workflows based on input type. Extracts structured information from documents, emails, reviews, feedback, and other sources.
development2
end_session
Canonical session close — commit, push, PR, release_task, reflection blocks, handover. Use /dump for emergency bail (no commit/PR/reflection).
data-ai2
Convert markdown documents to professionally formatted PDFs with academic-style typography, Roboto fonts, proper page layouts, and styling suitable for research documents, reviews, reports, and academic writing.
development2
diagram
Creating diagrams in any style — Mermaid flowcharts (structured, code-based) or Excalidraw (hand-drawn, organic). Use style parameter to select.
development2
daily
Daily note lifecycle — compose and maintain a factual daily note. Reports the state of the day; does not prioritise or recommend. SSoT for daily note structure.
data-ai2
survey
Survey a corpus, classify, and dispatch outputs. Three modes: retro (transcript review → issues), trend (longitudinal performance analysis), sweep (GitHub issue triage → fix-epics). Delegates execution to pauli (retro/trend) or jr (sweep) to keep main context clean.
data-ai2
verify
Judgement-based QA pass. Does this artifact meet its goal and serve its user? Demands excellence, not compliance. Owned by marsha; reads the spec's Fitness Rubric (designed upstream via /design-rubric).
development2
continue
Pause and hand back to the user with work still IN PROGRESS — emit a scannable resume summary and checkpoint the bound task, WITHOUT concluding. Use /end-session or /dump to finish the task completely.
testing2
task-lifecycle
The shared queue-to-execution spine for claiming work. Selects the next queued task, runs the premise + freshness gates, then either DISPATCHES it to a background surface (`/dispatch`) or CLAIMS and runs it INLINE in the current interactive session (`/pull`). Owns the select/gate/claim/verify/complete lifecycle so the two commands stay thin and never duplicate it. Invoked with a leading mode token: `dispatch: …` or `execute: …`.
testing2
strategic-review
Unified multi-agent review of any artifact — a document, plan, proposal, or pull request. The calling agent deploys rbg, pauli, and marsha in parallel, then @james reconciles their findings into one verdict. Pass `comment` and/or `fix` to write the result back to the review surface. Use `--critic` for a fast pauli-only pre-hoc critique.
testing2
dogfood
Delegated instruction testing — write instructions, commission contextless execution, observe friction, iterate, review quality, codify.
testing2
end_session
Canonical session close — commit, push, PR, release_task, reflection blocks, handover. Use /dump for emergency bail (no commit/PR/reflection).
data-ai2
task-lifecycle
The shared queue-to-execution spine for claiming work. Selects the next queued task, runs the premise + freshness gates, then either DISPATCHES it to a background surface (`/dispatch`) or CLAIMS and runs it INLINE in the current interactive session (`/pull`). Owns the select/gate/claim/verify/complete lifecycle so the two commands stay thin and never duplicate it. Invoked with a leading mode token: `dispatch: …` or `execute: …`.
testing2
daily
Daily note lifecycle — compose and maintain a factual daily note. Reports the state of the day; does not prioritise or recommend. SSoT for daily note structure.
data-ai2
supervisor
The single authoritative supervision process for any delegate-and-verify work — at every scale: one epic, a release spanning many epics (portfolio), or conversational orchestration of background workers (`/goal` "don't get involved yourself, make sure it gets done", `/dogfood`). Stateless tick driven by `/loop`; cross-tick state lives in the task body. Any orchestrator MUST invoke this skill for supervision; never hand-roll it inline.
development2
remember
Unified memory skill: immediate mode (/remember) persists knowledge via PKB MCP; maintenance mode (/sleep, GHA cron) runs periodic consolidation — transcript mining, knowledge synthesis, data quality, brain sync.
tools2
planner
Strategic planning agent — graph structure ownership, task decomposition, knowledge-building, and PKM maintenance. Works on WHAT exists and HOW it relates.
development2
strategic-review
Unified multi-agent review of any artifact — a document, plan, proposal, or pull request. The calling agent deploys rbg, pauli, and marsha in parallel, then @james reconciles their findings into one verdict. Pass `comment` and/or `fix` to write the result back to the review surface. Use `--critic` for a fast pauli-only pre-hoc critique.
testing2
verify
Judgement-based QA pass. Does this artifact meet its goal and serve its user? Demands excellence, not compliance. Owned by marsha; reads the spec's Fitness Rubric (designed upstream via /design-rubric).
development2
dbt
dbt (data build tool) implementation of the analyst transformation layer. Use when a project has a dbt/ directory or you need to build, test, or document SQL transformations as version-controlled, reproducible dbt models. This is the dbt-specific HOW for the tech-agnostic principles in the aops-tools analyst skill.
tools2
narrative-digest
Launder supervisor/worker task-log output into a Nic-facing narrative — what happened, where things are headed, and what (if anything) is genuinely his to decide. Never relays raw process detail (worker IDs, thread pointers, log paths) or verbatim task-log stream-of-consciousness.
testing2
dump
Emergency session bail — fast resume task + short handover, no commit/PR/reflection. For when you (or the user) need a clean context now. Use /end-session for canonical close.
data-ai2
end_session
Canonical session close — commit, push, PR, release_task, reflection blocks, handover. Use /dump for emergency bail (no commit/PR/reflection).
data-ai2
streamlit
Streamlit implementation of the analyst presentation layer. Use when building or updating a Streamlit dashboard that displays pre-computed research data. This is the Streamlit-specific HOW for the tech-agnostic principles in the aops-tools analyst skill — display only, never transform.
tools2
planner
Strategic planning agent — graph structure ownership, task decomposition, knowledge-building, and PKM maintenance. Works on WHAT exists and HOW it relates.
development2
project
Scaffold research project repositories with smart defaults — repo creation, directory structure, CI/CD, documentation, and PKB integration in one pass.
testing2
peer-review
Peer review of research funding applications and academic submissions. Scheme-agnostic — fetches current criteria from the relevant handbook each round, since weights and language change. Covers Detailed Assessor and College-of-Experts / General Assessor roles, plus collegial draft review.
data-ai2
aops
Core academicOps skill — institutional memory, strategic coordination, workflow routing, and framework governance. Merges butler (chief-of-staff) with framework development conventions.
development2
craft
Instruction quality gate — reviews agent instructions (task bodies, workflow steps, skill procedures, self-test protocols) for shallow-execution vulnerabilities before deployment. Two modes: author (pre-hoc review) and audit (trace a failure back to the instruction gap). The bar is excellence, not compliance.
testing2
program
Program / portfolio supervision — the autonomous top loop above /supervisor. "Ready the release" → discover and decompose the constituent epics → run /supervisor on each → surface only escalations + merge-ready PRs. Stateless tick driven by /loop; all cross-tick state lives in the program task body.
tools1
peer-review
Peer review of research funding applications and academic submissions. Scheme-agnostic — fetches current criteria from the relevant handbook each round, since weights and language change. Covers Detailed Assessor and College-of-Experts / General Assessor roles, plus collegial draft review.
data-ai1
design-rubric
Design-stage fitness rubric — persona immersion, scenario design, dimensions that define what excellence looks like for the people a feature serves. Two modes — author (produce a rubric for a new spec) and critique (red-team an existing spec). Output lives on the spec, not in the verification brief. Owned by pauli.
content-media1
review
Assist the user in reviewing academic work (papers, dissertations, drafts). Focuses on preparation, navigation, and synthesis, NOT replacing critical judgment.
testing
cowork-sync
Mirror PKB tasks onto the Cowork native task list at claim time and sync completion back to PKB. Cowork-only; ships only in the cowork build of aops-core.
development
active-loop
Iterative improvement protocol — declare a measurable target, run cycles via /loop, experiment, measure, learn, and accumulate work in a DRAFT PR.
tools
research
Academic research methodology guardian. Ensures agents working on empirical research maintain methodological integrity: research questions drive all design decisions, methods are appropriate and justified, data collection quality is verified before proceeding, and convenience shortcuts that compromise validity are caught and refused.
testing
annotations
Scan and process inline HTML comments for human-agent collaboration. Finds <!-- @nic: --> or <!-- @ns: --> comments and responds with dated <!-- @claude: --> replies. Works on markdown, Python, and other text files.
development
briefing-bundle
Generate morning briefing bundle with decision coversheets, email drafts, and annotation targets from the daily note. Run /daily first.
testing
decision-extract
Extract pending decisions from task queue, prioritize by blocking count, output to daily note for batch processing.
testing
extractor
Archive information extraction - assess archival documents and identify information worth preserving in the knowledge base.
development
garden
Incremental PKM maintenance - weeding, pruning, linking, consolidating. Tends the knowledge base bit by bit.
data-ai
python-dev
Write production-quality Python code following fail-fast philosophy, type safety, and modern best practices. Enforces rigorous standards for academic and research code where correctness and replicability are paramount.
development
training-set-builder
Extract structured training examples from document sets to build datasets for teaching LLMs specific tasks or styles. Use when processing review documents, feedback annotations, or revision histories.
development
convert-to-md
Batch convert documents (DOCX, PDF, XLSX, TXT, PPTX, MSG, DOC) to markdown, preserving tracked changes and comments.
documentation
sleep
Periodic consolidation agent — unified into the /remember skill (maintenance mode). This stub exists for backwards compatibility with installed GHA workflows.
data-ai
excalidraw
Creating visually compelling, hand-drawn diagrams with organic mind-map layouts and accessibility-focused design.
content-media
flowchart
Creating clear, readable, and attractive Mermaid flowcharts with best practices for accessibility, layout, and maintainability.
data-ai
debug-headless
Debug Claude Code or Gemini CLI in headless mode with full output capture.
tools
aops-tools/skills/style
Analyze writing samples and create a comprehensive personal writing style guide
tools
hdr
HDR (Higher Degree Research) student task conventions, reference letter workflows, and document access patterns.
documentation
dump
Emergency session bail — fast resume task + short handover, no commit/PR/reflection. For when you (or the user) need a clean context now. Use /end-session for canonical close.
data-ai
qa
QA verification, qualitative assessment, criteria design, and test planning
testing
assess-hydrator
Assess hydrator quality using real session data
testing
audit
Framework index curation and acceptance testing. Ensures documentation stays in sync with implementation.
development
burst-supervisor
Long-running iterative supervisor — dispatch work items to polecat workers across multiple bursts with state recovery.
tools
decision-apply
Process annotated decisions from daily note, update task statuses, and unblock dependent tasks.
testing
email-triage
Email triage workflow with mandatory archive receipt logging to task body
data-ai
dogfood
Generic reflective execution loop — learn from doing, capture friction, improve instructions
tools
eval
Agent session quality assessment — merged into /qa as Agent Session Evaluation mode
testing
fact-check
Verify factual claims in documents against authoritative sources. Catches hallucinations, fabricated quotes, and misattributed claims.
testing
osb-drafting
Generate draft OSB case decisions with IRAC analysis, position variants, and precedent support. Use when preparing case analysis for Oversight Board deliberation.
testing
process-bundle
Process annotated briefing bundle — execute decisions, stage email drafts, create tasks from annotations. Never auto-sends email.
testing
ground-truth
Establish and refine ground truth labels for evaluation datasets. Use when creating, reviewing, or updating labels for any judgment/reasoning task.
data-ai
recap
Reconstruct plain-English narrative of recent work from session summaries
data-ai
skill
--- name: skill title: Skill category: instruction --- # Review Training Data Extraction Skill Extract training pairs (review feedback → source evidence) from matched peer review/source document pairs to build a dataset for teaching LLMs to perform academic peer review in Nic's style. ## Purpose Process matched review/source pairs to create training data that captures: 1. **Review feedback units** - specific comments, suggestions, critiques 2. **Source evidence** - the text/pattern in the s
development
session-insights
Generate comprehensive session insights from transcripts using a Claude subagent
data-ai
framework
Project-local framework development skill — workflow routing, task lifecycle, and categorical conventions for working on academicOps
development