「ols」の検索結果

Expert in Cursor AI IDE - the leading AI-powered code editor. Covers Rules files for project-specific AI behavior, Plan Mode for structured development, Background Agents for parallel work, and advanced features like voice mode and browser integration. Essential for developers maximizing AI-assisted coding productivity. Use when "cursor ai, cursor rules, .cursorrules, cursor composer, cursor agent, plan mode, background agent, cursor tips, cursor, ide, ai-coding, editor, productivity, agent, developer-tools" mentioned.

日本語の概要は準備中です。原文の説明を表示しています。

DeFi protocol specialist for AMMs, lending protocols, yield strategies, and economic securityUse when "defi, amm, liquidity pool, lending protocol, yield farming, oracle, liquidation, tokenomics, tvl, impermanent loss, defi, amm, lending, yield, liquidity, oracle, tokenomics, protocol, ethereum, web3" mentioned.

日本語の概要は準備中です。原文の説明を表示しています。

Use when reviewing contracts, extracting key terms, identifying risks, or building contract analysis tools - covers NLP approaches, clause identification, and risk scoringUse when ", " mentioned.

日本語の概要は準備中です。原文の説明を表示しています。

Autonomous agents are AI systems that can independently decompose goals, plan actions, execute tools, and self-correct without constant human guidance. The challenge isn't making them capable - it's making them reliable. Every extra decision multiplies failure probability. This skill covers agent loops (ReAct, Plan-Execute), goal decomposition, reflection patterns, and production reliability. Key insight: compounding error rates kill autonomous agents. A 95% success rate per step drops to 60% by step 10. Build for reliability first, autonomy second. 2025 lesson: The winners are constrained, domain-specific agents with clear boundaries, not "autonomous everything." Treat AI outputs as proposals, not truth. Use when "autonomous agent, autogpt, babyagi, self-prompting, goal decomposition, react pattern, agent loop, self-correcting agent, reflection agent, langgraph, agentic ai, agent planning, autonomous, agents, langgraph, react, planning, reflection, guardrails, reliability, checkpointing" mentioned.

日本語の概要は準備中です。原文の説明を表示しています。

Expert in Claude Code hooks - user-defined shell commands that execute at specific points in Claude Code's lifecycle. Provides guaranteed automation that doesn't rely on the LLM "remembering" to do something. Essential for enterprise workflows, code quality enforcement, and deterministic behavior in agentic coding. Use when "claude code hooks, automate claude, block tool use, pre tool use, post tool use, session start hook, claude lifecycle, claude-code, automation, hooks, lifecycle, cli, workflow, devtools" mentioned.

日本語の概要は準備中です。原文の説明を表示しています。

The enhancement layer for AI-generated content. This skill covers AI-powered visual effects, compositing, upscaling, restoration, and post-production magic—turning raw AI output into polished, professional content. AI generation gets you 80% of the way. Visual effects get you the remaining 20% that separates "clearly AI" from "how did they do that?" This skill covers ComfyUI workflows, Runway's AI tools, intelligent upscaling, rotoscoping, color grading, and the integration of AI elements into traditional footage. The practitioners of this skill are technical artists who understand both traditional VFX workflows and the new AI-native approaches that are revolutionizing post-production. Use when "AI visual effects, VFX, upscale, upscaling, composite, rotoscope, background removal, color grade, inpaint, outpaint, style transfer, enhance, ComfyUI, post-production AI, vfx, visual-effects, compositing, upscaling, post-production, comfyui, enhancement, color-grading" mentioned.

日本語の概要は準備中です。原文の説明を表示しています。

The intersection of AI generation and performance marketing. This skill covers creating ad creatives at scale using AI tools—from static images to video ads to dynamic creative optimization—while maintaining the conversion focus that performance marketing demands. Traditional creative testing is slow: create 5 variants, wait weeks for data, iterate. AI-powered ad creative is fast: generate 50 variants in an hour, test simultaneously, learn in days, iterate continuously. The constraint is no longer production capacity—it's testing velocity and creative strategy. This skill bridges AI generation capability with advertising effectiveness, ensuring that AI-created ads don't just look good—they convert. Use when "AI ads, ad creative, performance creative, ad generation, creative testing, ad variants, DCO, dynamic creative, Meta ads, Google ads, ad fatigue, conversion creative, advertising, performance-marketing, creative-testing, ai-ads, conversion, paid-media, scale" mentioned.

日本語の概要は準備中です。原文の説明を表示しています。

The systematic orchestration of AI-powered marketing workflows that combine content generation, approval processes, multi-channel distribution, and quality gates into cohesive automation systems. This skill integrates AI generation tools (Jasper, Claude, GPT) with automation platforms (Zapier, Make, n8n) and marketing systems to build scalable content pipelines. It focuses on maintaining brand consistency, implementing rigorous quality gates, and balancing automation with strategic human oversight. Key capabilities include designing parallel approval flows, monitoring costs, and architecting "invisible" automation that enhances productivity without sacrificing quality.Use when "AI workflow, automate content, content automation, workflow automation, AI pipeline, automated marketing, content distribution automation, approval workflow, scale content production, AI orchestration, automation, workflow, ai-orchestration, content-pipeline, approval-workflow, multi-channel, quality-gates, cost-control" mentioned.

日本語の概要は準備中です。原文の説明を表示しています。

Build comprehensive AI-native brand asset systems that maintain consistency across all AI-generated content. Train AI tools on brand guidelines, create reusable prompt libraries, and manage visual/voice assets at scale. Use when ", " mentioned.

日本語の概要は準備中です。原文の説明を表示しています。

Expert in building products that wrap AI APIs (OpenAI, Anthropic, etc.) into focused tools people will pay for. Not just "ChatGPT but different" - products that solve specific problems with AI. Covers prompt engineering for products, cost management, rate limiting, and building defensible AI businesses. Use when "AI wrapper, GPT product, AI tool, wrap AI, AI SaaS, Claude API product, " mentioned.

日本語の概要は準備中です。原文の説明を表示しています。

Comprehensive patterns for building AI-powered code generation tools, code assistants, automated refactoring, code review, and structured output generation using LLMs with function calling and tool use. Use when "code generation, AI code assistant, function calling, structured output, code review AI, automated refactoring, tool use, code completion, agent code, " mentioned.

日本語の概要は準備中です。原文の説明を表示しています。

The new frontier of audio: AI-generated music with Suno and Udio, AI sound effects with ElevenLabs, AI voice cloning, and AI audio enhancement. This skill covers the full spectrum of neural audio synthesis and manipulation. We're in an era where a single person can produce broadcast-quality audio content that would have required studios, musicians, and engineers. AI doesn't replace musical talent—it democratizes production capability. Original compositions in minutes. Custom sound design in seconds. Voice in any language instantly. The practitioners of this skill understand both the creative potential and the ethical considerations. They know when AI audio enhances creativity and when it needs human touch. They're audio directors who orchestrate AI tools like instruments in a symphony. Use when "AI music, generate music, Suno, Udio, AI audio, AI sound, generate soundtrack, background music, sound effects, audio generation, jingle, AI score, stem separation, ai-audio, music, suno, udio, sound-effects, soundtrack, generation, audio-branding, production" mentioned.

日本語の概要は準備中です。原文の説明を表示しています。

The specialized craft of writing copy that sells—headlines that stop thumbs, body copy that persuades, CTAs that convert. This is copywriting specifically tuned for advertising: paid media, performance marketing, direct response. Ad copy is different from content writing. Every word costs money (CPM/CPC). Every character competes for attention. There's no time for warm-up—you have one chance to hook, persuade, and convert. The best ad copy feels effortless but is engineered word by word. In the age of AI-generated creative at scale, ad copywriting becomes both more important (you need more variations) and more challenging (everyone has access to the same tools). The differentiation is in the strategy, the insight, and the craft that AI assists but doesn't replace. Use when "ad copy, headline, ad text, CTA, call to action, Facebook ad, Google ad, ad variations, direct response, performance copy, primary text, ad headline, ad-copy, headlines, direct-response, performance, conversion, cta, advertising" mentioned.

日本語の概要は準備中です。原文の説明を表示しています。

Transparent, itemized impressions of AI-generated WRITING STYLE — the repo's ONLY non-integrity track. Two passes: a deterministic defensive-hedge density screen (tools/check_ai_style.py, AIS-DEFENSIVE-HEDGE) plus a fresh cross-model GROSS-cases-only semantic pass over the 13 AIS-* style tells (broken narrative arc, LLM phrase tics, jargon-stuffing, invented codenames, clause/formula walls, gratuitous pseudocode, bullet overuse, bold-module spam, restatement loops, focus drift, single-style figures, appendix dumping). Every finding is named, LOCATED, span-anchored to the evidence ledger (claims.json), carries not_integrity_finding:true + false_positive_risk:high + an fp_case, and gets ZERO verdict weight: the adjudicator forces it to info, excludes it from overall_verdict, and renders it in a SEPARATE report section. NOT an AI-text classifier — no scores, no "this is AI-written", no authorship probability; a paper can be CLEAN_GIVEN_EVIDENCE and still list many. Emits ai-style-impressions.findings.json; computes NO verdict; if a tell is actually substantive it routes to the integrity auditor. Triggers: "AI style", "vibe check", "writing fingerprint", "AI 文风", "vibe paper".

日本語の概要は準備中です。原文の説明を表示しています。

Checkable-ish surface presentation signals a reviewer notices first — duplicate/near-identical tables, leftover pipeline/template strings, too-few or LLM-looking figures, and page-padding. AUXILIARY ONLY and weak by design: a deterministic pass (tools/check_presentation.py — dup-table + pipeline-artifact) plus a fresh cross-model GROSS-cases-only semantic pass (thin-float, LLM-figure, page-padding), every above-info finding span-anchored to the evidence ledger (claims.json). The adjudicator CAPS everything at minor (SURFACE_ONLY_SKILLS + SURFACE_PATTERNS) — these contribute at most SOFT_FLAGS, never a HARD verdict — default false_positive_risk:high. NOTE: the pure AI writing-STYLE impressions (AI-flavor prose, defensive 'not-X-but-Y' hedging, narrative-arc, jargon-stuffing, invented codenames) MOVED to the zero-verdict-weight AIS track — for those use skills/ai-style-impressions, NOT this. Emits presentation-signals.findings.json; NEVER computes the verdict. Triggers: "presentation signals", "surface check", "duplicate tables", "排版信号".

日本語の概要は準備中です。原文の説明を表示しています。

Audit experiment integrity against the evidence ledger. At L2 (repo + result files present) a fresh cross-model reviewer reads the eval code line-by-line for fake/derived ground truth, score self-normalization, phantom results (a paper number with no backing file/key), dead/uncalled metric code, verified-scope inflation, method-described ≠ method-evaluated drift, synthesized-looking results, placeholder/fake data still wired into a released result, code-output ≠ reported-number mismatch, and missing reproducibility artifacts (an empirical/agent/LLM paper shipping neither code nor the prompts/configs its results need) — every finding span-anchored to a ledger claim_id. At L0/L1 (PDF / source only) the same patterns are surfaced as info-level 'could-not-verify' signals where the ledger gives an anchor (observability_level_required:2) — NEVER a fraud verdict from a PDF. The reviewer PROPOSES findings; tools/adjudicate_findings.py computes the verdict. Detect-only. Triggers: "experiment forensics", "audit the results", "check the eval code", "实验诚实度".

日本語の概要は準備中です。原文の説明を表示しています。

End-to-end substantive-integrity forensic sweep of a research paper (especially autoresearch / AI-Scientist-style output). Orchestrates the whole pipeline: ingest (arxiv-id | pdf | dir → working dir + pdftotext for L0) → /evidence-ledger (artifact manifest + observability level L0/L1/L2 + span-anchored claims.json) → fan out the integrity auditor skills (consistency, citation, baseline, experiment, presentation, proof-derivation, eval-design — each reads the ledger, emits span-anchored findings) + the zero-verdict-weight AIS writing-style track → advisory memos (/adversarial-case-builder + /novelty-duplication-advisory, no verdict weight) → deterministic tools/adjudicate_findings.py (--ledger REQUIRED) → reviewer-ready Integrity Forensics Report. Cross-model (fresh codex per dimension) and reviewer≠adjudicator: the model proposes findings, the deterministic adjudicator decides the verdict. Observability-aware, detect-only, never an opaque AI-text classifier (a separate zero-weight AIS section lists AI writing-style impressions, never moving the verdict). Triggers: "anti-autoresearch", "integrity audit this paper", "forensic review", "audit a submission", "审一篇投稿的诚信".

日本語の概要は準備中です。原文の説明を表示しています。

Audit whether a paper's EVALUATION DESIGN actually measures what it claims and whether its reporting is complete — the validity layer family D (experiment-forensics) cannot reach. Three patterns: train/test leakage means the reported score may not measure generalization (HP-EVAL-LEAKAGE — adopts the Kapoor & Narayanan 8-type / 3-category leakage taxonomy; the illegitimate-proxy / sampling-bias / pretraining-contamination subtypes hand off as needs_external_check, naming but NEVER running Oren-2023 exchangeability / Shi-2023 Min-K% / Golchin-2023 Time-Travel / BIG-bench canary); a load-bearing LLM judge is conflicted (same model/family as a compared system) or unvalidated (no human-agreement, no bias control) (HP-JUDGE-VALIDITY); a declared condition/metric is dropped or switched to favor the method, or 'best' is chosen with no held-out set (HP-SELECTIVE-REPORTING). Verdict-bearing at L0/L1 from the DESCRIBED protocol — NOT repo-gated like experiment-forensics; L2 only CONFIRMS against split/preprocessing/result files. A fresh cross-model reviewer (gpt-5.6-sol xhigh, read-only, fresh thread per pass) PROPOSES findings, each span-anchored to a ledger claim_id; tools/adjudicate_findings.py DECIDES the verdict. Leakage and under-reporting are usually HONEST methodological errors — every finding describes a discrepancy to CHECK, never an accusation. An LLM generating GROUND-TRUTH labels is HP-FAKE-GT (experiment-forensics) — routed there, not here. Emits eval-design-forensics.findings.json; computes NO verdict. Detect-only. Triggers: "eval design audit", "evaluation validity", "train/test leakage", "data leakage", "is the score measuring generalization", "LLM judge bias", "is the judge validated", "selective reporting", "cherry-picked results", "评估设计审计", "评测有效性", "数据泄漏", "训练测试集泄漏", "裁判模型有没有验证", "选择性报告".

日本語の概要は準備中です。原文の説明を表示しています。

Audit whether a paper's baseline comparisons are COMPLETE, FAIR, and SIGNIFICANT: a required recent SOTA baseline is missing while 'best/SOTA' is claimed (HP-MISSING-BASELINE); a baseline is undertuned / given less compute-tuning-data, run at a mismatched config, or the equal-budget ablation-as-baseline is absent (HP-WEAK-BASELINE); 'outperforms' is asserted over overlapping error bars or with no variance/seeds (HP-SIG-OVERLAP); and a cross-row 'improves over baseline by X%' is arithmetically wrong (HP-DELTA-ERROR, cross-row form only). A versioned per-domain baseline profile + a live leaderboard/recency search are assembled by the EXECUTOR as structured facts; a fresh cross-model reviewer (gpt-5.6-sol xhigh, read-only, fresh thread per dimension) PROPOSES findings, each span-anchored to a ledger claim_id; tools/adjudicate_findings.py DECIDES the verdict. Works at L0 (stated comparisons) and deepens at L2 (configs/result files). A completeness question it cannot settle internally becomes needs_external_check, never a guessed missing baseline. Emits baseline-comparison-audit.findings.json; computes NO verdict. Detect-only. Triggers: "baseline audit", "missing baselines", "is the comparison fair", "weak baseline", "baseline 误报", "SOTA earned?".

日本語の概要は準備中です。原文の説明を表示しています。

Synthesize the single strongest EVIDENCE-BOUND reviewer case to reject a paper, built ONLY from the evidence ledger (claims.json) + the other auditors' confirmed findings — never free-floating LLM critique. Two fresh cross-model codex threads: an attack writes the ~200-word rejection paragraph (every accusation tagged to an existing claim_id/finding_id), a defense decomposes it and rules each point against the anchored evidence. MEMO-ONLY: emits adversarial-case-builder.memo.md (fed to the adjudicator via --memo) and carries NO verdict weight — tools/adjudicate_findings.py lists it in ZERO_WEIGHT_SKILLS and caps it at info. Honest-null allowed (the paper may survive). Run LAST. Detect-only. Adapted from ARIS kill-argument. Triggers: "adversarial case", "strongest objection", "rejection memo", "kill argument", "最强拒稿点".

日本語の概要は準備中です。原文の説明を表示しています。

MEMO-ONLY prior-work overlap advisory: surfaces the two ADVISORY taxonomy signals neither a tool nor a model can decide from the paper alone — ADV-TRIVIAL-COMBINATION (standard A+B+C / 缝合 stapling) and ADV-DUPLICATE-PUBLICATION (repackaged / duplicate submission). The executor RETRIEVES candidate prior work (DBLP fuzzy-title + boolean · WebSearch · WebFetch) from the paper's own title + contribution spans in the evidence ledger; TWO fresh cross-model codex reviewers (one per axis) LAY OUT the overlap side-by-side against each anchored contribution claim. It NEVER rules 'trivial' or 'duplicate' (that is a human judgment) and absence of a match is NOT evidence of originality. Emits novelty-duplication-advisory.memo.md + an info-only findings mirror; carries NO verdict weight — tools/adjudicate_findings.py lists it in ZERO_WEIGHT_SKILLS and caps it at info. Detect-only. Adapted from ARIS novelty-check, reframed from 'is MY idea novel' to 'here is the overlap a reviewer should weigh'. Triggers: "novelty advisory", "duplication check", "prior-work overlap", "is this stapling", "缝合", "查重", "重复发表", "duplicate submission".

日本語の概要は準備中です。原文の説明を表示しています。

Writes a handover document so another person, another agent, or a different tool can pick up multi-session work cold. Use when handing work to someone else, switching tools, or pausing work that will not resume in the same session.

日本語の概要は準備中です。原文の説明を表示しています。

Refuses "flaky" as a diagnosis. Turns an intermittent failure into a reproduction by isolating it from the environment, naming the variable that controls it, forcing that variable, and measuring before and after with enough runs to count. Separates real defects from environment contention with evidence, and never retries to green or widens a threshold. Use when a test, performance check, or deploy step fails intermittently or passes on rerun.

日本語の概要は準備中です。原文の説明を表示しています。

AWS Advisor workflow skill. Use this skill when the user needs Expert AWS Cloud Advisor for architecture design, security review, and implementation guidance. Leverages AWS MCP tools for accurate, documentation-backed answers. Use when user asks about AWS architecture, security, service selection, migrations, troubleshooting, or learning AWS. Triggers on AWS, Lambda, S3, EC2, ECS, EKS, DynamoDB, RDS, CloudFormation, CDK, Terraform, Serverless, SAM, IAM, VPC, API Gateway, or any AWS service. Do NOT use for non-AWS cloud providers or general infrastructure without AWS context and the operator should preserve the upstream workflow, copied support files, and provenance before merging or handing off.

日本語の概要は準備中です。原文の説明を表示しています。