「long context」の検索結果

Manage LLM context budgets—prioritization, summarization, compaction, and what to load vs reference. Use for long sessions, large repos, or multi-doc tasks.

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。

Re-inject Synapse A2A initial instructions after context has been cleared. Use this skill when the agent has lost its identity, A2A communication protocol, or configuration after /clear or context reset, when synapse send/reply commands stop working, or when the agent no longer recognizes its SYNAPSE_AGENT_ID. Triggered by /synapse-reinst command.

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。

Compact current session context into a handoff document for seamless continuation in a new session. Use when context window is exhausted, switching machines, handing work to another agent, or ending a long session mid-task.

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。

Full-funnel paid advertising operations for B2B SaaS. Use this skill for PPC strategy, Google Ads optimization, LinkedIn Ads, social media advertising, programmatic buying, creative strategy, attribution modeling, budget allocation, budget pacing and planned-vs-delivered reconciliation, audience overlap and campaign self-competition audits, suppression lists, ROAS improvement, media spend optimization, and the media no auction sells — newsletter and podcast sponsorships, community and industry-publication placements, paid review-site listings on G2 and Capterra, and pay-per-lead content syndication. It also covers buying on the AI answer surface — ads inside Google AI Overviews and AI Mode that existing Search, Shopping and Performance Max campaigns already serve into with no opt-out, no placement targeting and no segmented reporting, and ChatGPT Ads and Microsoft Copilot as separate opt-in platforms with their own plan-tier reach limits, no negative keywords, and a landing-page crawler gate. It also owns the conversion signal the bidding algorithms learn from — CRM pipeline stages (SQL, opportunity, closed-won) sent back to Google Ads, LinkedIn, Meta and Microsoft Ads through offline conversion import, enhanced conversions for leads, the Google Data Manager API and the LinkedIn Conversions API, with click-ID capture, stage values, dedupe, EEA consent signals and retractions. It also gates connected TV (CTV) and streaming video for B2B — whether a reach medium belongs in the plan, how a vendor ties a TV household to a target account (member data, IP-to-account graph or contextual), the server-side ad insertion (SSAI) fraud questions, Programmatic Guaranteed deals, and a holdout lift test because CTV has no click. It also runs person-led paid social — LinkedIn Thought Leader Ads and Meta partnership ads that put spend behind an employee's or creator's post — on a paid-usage term, an in-post disclosure check and an end-date register, because the platform approval carries no term. Also triggers on: CTV, connected TV, OTT, streaming TV ads, should we run CTV, LinkedIn CTV, B2B CTV, account-based CTV, SSAI fraud, thought leader ads, boost an employee's post, sponsor a creator's post, partnership ads, whitelisting, creator usage rights, PPC, Google Ads, LinkedIn Ads, social ads, programmatic, ad creative, attribution, media budget, paid campaigns, ROAS, CPA, ad spend, newsletter sponsorship, podcast sponsorship, community sponsorship, sponsor a newsletter, content syndication, pay-per-lead, cost per lead vendor, media kit, rate card, insertion order, direct buy, publisher partnership, make-good, G2 paid listing, Capterra ads, review site advertising, sponsorship ROI, budget pacing, we underspent our budget, are our campaigns competing with each other, audience overlap, self-competition, campaign cannibalization, suppression list, exclusion list, ad frequency across channels, how many ad accounts do we have, ads in AI Overviews, AI Mode ads, ChatGPT Ads, OpenAI Ads Manager, Copilot ads, conversational ads, should we advertise on ChatGPT, can we buy our way into the AI answer, our ads are showing in AI Overviews and we cannot report on it, context hints, OAI-AdsBot, our ad landing page fails review, robots.txt is blocking our ads, our paid leads are junk, Google Ads is optimizing for the wrong leads, offline conversion import, OCI, enhanced conversions for leads, send CRM stages back to Google Ads, upload SQLs to LinkedIn, LinkedIn Conversions API, LinkedIn CAPI, Meta conversion leads, CRM to ad platform, gclid, li_fat_id, msclkid, value-based bidding, optimize for pipeline not form fills, low match rate, conversion lag.

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。

Keep long conversations coherent — summarization checkpoints, entity tracking, topic stacks, and graceful context handoffs. Use when dialogues run long or span sessions and the agent keeps losing the thread.

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。

Manage LLM context windows deliberately — budgeting, prioritization, summarization, and retrieval instead of dumping everything in. Use when prompts grow long, costs rise, or the model loses track of key details.

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。

Add /compact command for manual context compaction. Solves context rot in long sessions by forwarding the SDK's built-in /compact slash command. Main-group or trusted sender only.

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。

Guide for implementing and maintaining context as a managed artifact alongside code, enabling consistent AI interactions and team alignment through structured project documentation.

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。

Capture, gate, query, and render long-term project knowledge through the gated `mewkit wiki` subsystem. Use to create a wiki, propose/approve candidates (scanner-gated), hand off a skill's terminal artifact as a scanned candidate, recall context, search the FTS index, or list pages. Agents may only PROPOSE candidates; canonical pages are written only via human approve. NOT for short-lived JSON memory (see mk:memory); NOT for fetching external sources (see mk:wiki-research); NOT for HTML snapshots (see mk:wiki-render).

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。

Otimiza atenção em transformers com Flash Attention para ganho de 2-4x em velocidade e redução de 10-20x em memória. Use ao treinar/executar transformers com sequências longas (>512 tokens), ao encontrar problemas de memória GPU com atenção, ou quando precisa de inferência mais rápida. Suporta SDPA nativo do PyTorch, biblioteca flash-attn, H100 FP8 e sliding window attention.

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。

[Utilities] Use when managing context window usage, compressing long sessions, or optimizing token usage.

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。

Add /compact command for manual context compaction. Solves context rot in long sessions by forwarding the SDK's built-in /compact slash command. Main-group or trusted sender only.

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。

Add /compact command for manual context compaction. Solves context rot in long sessions by forwarding the SDK's built-in /compact slash command. Main-group or trusted sender only.

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。

Apply when context is filling up: large outputs, long files, repeated reads, fan-out planning. Route bulk to subagents; keep summaries in the main thread, not raw payloads.

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。

This skill helps agents use Figma's use_figma MCP tool in the Slides context. Can be used alongside figma-use which has foundational context for using the use_figma tool.

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。

This skill helps agents use Figma's use_figma MCP tool in the FigJam context. Can be used alongside figma-use which has foundational context for using the use_figma tool.

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。

Three-file state management for session continuity. Maintains STATE.md, DECISIONS.md, and PROGRESS.md as human-readable session context alongside the SQLite ticket database. Use this skill when updating workflow state, logging decisions, tracking session progress, or resuming work after a context break.

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。

Agent context budget guard and artifact compression workflow. USE FOR: image-heavy QA, compare board review, screenshot batches, large keep/todo/task manifests, long notes handoff, and md/json diff summarization. Trigger this before reading or forwarding heavy artifacts when token growth is a risk.

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。

Extract exactly what the person typed in a Claude Code session (script, no agent context spent), have a fresh agent write a faithful structured summary under it, and optionally have an adversarial agent judge whether the working agent is still on track. Use at a checkpoint, before a handoff, after a compaction, when a long session may have drifted from what the owner asked, or when told to 'summarize what I said' or 'check we're on track'.

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。

Optimizes transformer attention with Flash Attention for 2-4x speedup and 10-20x memory reduction. Use when training/running transformers with long sequences (>512 tokens), encountering GPU memory issues with attention, or need faster inference. Supports PyTorch native SDPA, flash-attn library, H100 FP8, and sliding window attention.

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。

Persistent AI companion with long-term memory, project context awareness, and multi-session continuity across any LLM.

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。

Add /compact command for manual context compaction. Solves context rot in long sessions by forwarding the SDK's built-in /compact slash command. Main-group or trusted sender only.

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。

When to delegate to `cavecrew-investigator` (locate code), `cavecrew-builder` (1-2 file edit) or `cavecrew-reviewer` (diff review) instead of working inline or using `Explore`. Their output is compressed, so main context lasts longer.

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。

Implement persistent memory patterns for AI agents using AgentDB. Includes session memory, long-term storage, pattern learning, and context management. Use when building stateful agents, chat systems, or intelligent assistants.

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。