woosook0127
- 24 skills
- 0 followers
- 8 hours ago last updated
- ▌ Supervisor Intake · woosook0127Use when a new human request must be turned into Codex-native company workflow state: TASK, routing decision, acceptance criteria, branch intent, next team, and file-based handoff without using OMX.
- ▌ Implementation Plan · woosook0127Use when a code, model, automation, or documentation change needs an explicit Codex-native implementation plan with assumptions, target files, acceptance criteria, and verification before execution.
- ▌ Research Idea Screen · woosook0127Use when a research idea or hypothesis must be screened against literature, current repo overlap, existing baselines/runs/specs, novelty risks, comparison papers, and minimum credible experiment direction before coding.
- ▌ Experiment Feasibility · woosook0127Use when a proposed experiment needs a grounded GO, CONDITIONAL, or NO_GO decision covering backend fit, baseline preservation, data availability, compute burden, metric support, and validation path.
- ▌ Research To Experiment · woosook0127Use when a research workflow must connect paper reading, field gap analysis, idea formation, repo mapping, experiment design, execution, testing, result analysis, and skeptical closure.
- ▌ Literature Field Surveyor · woosook0127 bundleLiterature survey, paper search, unfamiliar field, research trend, field history, Q&A from papers: use when building, updating, or querying a paper-grounded Markdown knowledge base for an academic research field, including per-paper notes, synthesis files, and logged research Q&A.
- ▌ Run Test · woosook0127Use when code, scripts, configs, or documentation tooling need graduated verification commands: syntax, import, smoke, functional, integration, or regression checks, with read-only reporting of failures.
- ▌ Repo Audit · woosook0127Use when Codex must inspect the repository before architecture, refactor, experiment, automation, or model-work decisions; maps entry points, configs, datasets, engines, metrics, checkpoints, and reuse boundaries with file evidence.
- ▌ Audit Review · woosook0127Use when work needs final independent audit for policy compliance, provenance, baseline preservation, artifact completeness, result credibility, merge/promotion readiness, or caveated AUDIT_PASS/AUDIT_FAIL decision.
- ▌ Debug Review · woosook0127Use when develop output needs a blocking Codex-native debug quality gate: reproduce or smoke-test the claimed change, inspect logs/artifacts, verify branch/provenance when relevant, and issue PASS, PASS_WITH_WARNINGS, FAIL, or URGENT_BLOCKER.
- ▌ Execute Plan · woosook0127Use to carry out an approved or clearly implied implementation plan with surgical edits, progress tracking, baseline preservation, and handoff to verification.
- ▌ Final Report · woosook0127Use at the end of a planned task to produce a concise final report from the request, plan, edits, tests, artifacts, residual risks, and next actions.
- ▌ Paper Intake · woosook0127Use when a paper, citation, PDF, excerpt, or note set must be converted into a repo-grounded research or reproduction packet: executable claims, datasets, architecture, training recipe, evaluation protocol, ambiguities, and next experiment path.
- ▌ Skillset Sync · woosook0127Use when Codex must validate, synchronize, document, commit, or push the user's Codex skillset repository and keep project-local skill copies aligned without Claude/Notion-specific tooling.
- ▌ Artifact Audit · woosook0127Use for read-only audit of research notes, paper cards, plans, reports, drafts, and workflow artifacts for factual support, structure, cross-reference integrity, coverage gaps, and readiness before handoff.
- ▌ Paper Critique · woosook0127Use when a paper claim, reproduction plan, or paper-derived experiment needs skeptical review for weak baselines, unfair comparisons, leakage, missing ablations, under-specified protocol, or overclaimed gains.
- ▌ Run Experiment · woosook0127Use when a canonical or clearly defined experiment spec is ready to execute locally: resolve backend/config, record exact command, run least-invasive train/eval command, capture logs/artifacts, and report status or failure diagnostics.
- ▌ Skeptic Review · woosook0127Use when final experiment conclusions, result analyses, reproduction claims, or paper-derived claims need a disciplined skeptical pass for unfair baselines, leakage, tuning asymmetry, ignored regressions, and missing evidence.
- ▌ Analyze Project · woosook0127Use before substantial code, paper, or document work when Codex must map the given materials, detect the task mode, gather evidence, and produce a concise project analysis artifact before planning or execution.
- ▌ Analyze Results · woosook0127Use when completed experiment outputs, logs, metrics, predictions, checkpoints, or run summaries need interpretation against hypotheses, baselines, artifacts, regressions, and scientific claim status.
- ▌ Company Workflow · woosook0127 bundleUse when Codex must operate the user's company-style durable workflow without OMX: supervisor routing, research/develop/debug/audit team state, file-based handoffs, task documents, quality gates, baseline preservation, scientific closure, and Codex-native skill orchestration.
- ▌ Document Workflow · woosook0127Use for Codex-native document strategy, drafting, and refinement for reports, proposals, paper notes, rebuttals, reviews, documentation, and presentation scripts using local materials and versioned edits.
- ▌ Experiment Design · woosook0127Use when turning a user goal, paper packet, idea, or feasibility result into an executable experiment design: hypothesis, baseline, variant, ablations, metrics, artifacts, reproducibility metadata, and current-backend mapping.
- ▌ Framework Mapping · woosook0127Use when an idea, paper, model change, dataset, or experiment must be mapped onto the current repository backend, launcher/spec path, preserved baseline policy, and thinnest executable implementation route.