Results for “long context”

Decision guide for delegating to caveman-style subagents. Tells the main thread WHEN to spawn `cavecrew-investigator` (locate code), `cavecrew-builder` (1-2 file edit), or `cavecrew-reviewer` (diff review) instead of doing the work inline or using vanilla `Explore`. Subagent output is caveman-compressed so the tool-result injected back into main context is ~60% smaller — main context lasts longer across long sessions. Trigger: "delegate to subagent", "use cavecrew", "spawn investigator/builder/reviewer", "save context", "compressed agent output".

The English summary is being prepared. Showing the original description.

Context management specialist for multi-agent workflows and long-running tasks. Use when coordinating multiple agents, preserving context across sessions, or managing complex project state. Focuses on context compression and distribution.

The English summary is being prepared. Showing the original description.

Audits and reorganises what an AI coding agent carries in its context window: instruction files, path-scoped rules, skills, subagents, hooks and the running conversation. Use when someone says "the agent ignores our CLAUDE.md", "set this repo up for Claude Code / Codex / Cursor", "my AGENTS.md is too long", "quality drops in long sessions", "it keeps inventing APIs", "what should go in a skill versus the rules file", or asks how to hand work to a subagent. Grounded in the documented loading behaviour of Claude Code, with the equivalents for Codex, Gemini CLI and Cursor.

The English summary is being prepared. Showing the original description.

Delegate long-context analysis, research, and large-document summarization to Google Gemini CLI (1M-token context) via non-interactive `gemini -p`, preserving Claude context. Use when the user says 'use gemini'/'hand off to gemini' or a task needs 100K+ tokens of context.

The English summary is being prepared. Showing the original description.

Generate a state document (context transfer) that allows any fresh AI instance to resume work without conversation history. Use this skill whenever the user asks to checkpoint, create a state transfer, save state, generate a handoff document, create compaction insurance, or says anything like 'snapshot this', 'create a context transfer', 'checkpoint', 'save state for handoff', 'write a recovery doc', or 'transfer context'. Also trigger when the user mentions compaction concerns, session handoff, multi-session continuity, or wants to preserve the current conversation's state. Works for any conversation type: projects, research, debugging, investigations, design sessions, or long-running chats. This skill produces a structured markdown document optimized for AI cold-start recovery, not human notes.

The English summary is being prepared. Showing the original description.

AI internal coordination using stigmergic signals — managing information freshness in context and memory, decay rates for assumption staleness, and emergent coherent behavior from simple local protocols. Use during complex tasks where multiple sub-tasks must coordinate, when context has grown long and information freshness is uncertain, after context compression when information may have been lost, or when sub-task outputs need to feed into each other cleanly without degradation.

The English summary is being prepared. Showing the original description.

Decision guide for delegating to caveman-style subagents. Tells the main thread WHEN to spawn `cavecrew-investigator` (locate code), `cavecrew-builder` (1-2 file edit), or `cavecrew-reviewer` (diff review) instead of doing the work inline or using vanilla `Explore`. Subagent output is caveman-compressed so the tool-result injected back into main context is ~60% smaller — main context lasts longer across long sessions. Trigger: "delegate to subagent", "use cavecrew", "spawn investigator/builder/reviewer", "save context", "compressed agent output".

The English summary is being prepared. Showing the original description.

Creates comprehensive handoff documents for seamless AI agent session transfers. Triggered when: (1) user requests handoff/memory/context save, (2) context window approaches capacity, (3) major task milestone completed, (4) work session ending, (5) user says 'save state', 'create handoff', 'I need to pause', 'context is getting full', (6) resuming work with 'load handoff', 'resume from', 'continue where we left off'. Proactively suggests handoffs after substantial work (multiple file edits, complex debugging, architecture decisions). Solves long-running agent context exhaustion by enabling fresh agents to continue with zero ambiguity.

The English summary is being prepared. Showing the original description.

This skill should be used when the user asks to "compress context", "summarize conversation history", "implement compaction", "reduce token usage", or mentions context compression, structured summarization, tokens-per-task optimization, or long-running agent sessions exceeding context limits.

The English summary is being prepared. Showing the original description.

features.json を読み込み、pending な機能を1件ずつ tdd-guide エージェント（内部で tdd-cycle を使用）で実装するスキル。 Context anxiety（トークン上限接近による焦り）と Context rot（情報劣化）を防ぐため、 機能単位のループ・状態管理・implementation-notes.md への記録を担う。 builder エージェントが実装フェーズを開始するときは必ずこのスキルを呼ぶ。

The English summary is being prepared. Showing the original description.

Elite AI context engineering specialist mastering dynamic context management, vector databases, knowledge graphs, and intelligent memory systems. Orchestrates context across multi-agent workflows, enterprise AI systems, and long-running projects with 2024/2025 best practices. Use PROACTIVELY for complex AI orchestration.

The English summary is being prepared. Showing the original description.

Organize existing project documentation into rules, structure, current status, and decision history. Connect agents to authoritative sources and define update rules to address outdated documents, repeated context discovery, and accidental recreation of intentionally removed files.

• Organizing fragmented project documentation
• Preserving context across agent sessions
• Recording why files or approaches were intentionally removed

Integrate Flash Attention into Transformer models to speed up attention, the calculation that relates input elements, while reducing GPU memory use. Covers PyTorch integration, performance measurement, output comparison, and a separate H100 optimization path.

• Optimize training with long input sequences
• Reduce memory use during long-context inference
• Measure speed and compare model outputs

Organize the information an AI coding agent needs for each task. This skill guides project rules, relevant file selection, conversation summaries, and session handoffs so the agent can follow existing conventions and retain key decisions.

• Set up project rules for AI coding agents
• Select relevant files before starting a task
• Trim stale information from a long conversation

Design AI agents that break down goals, plan actions, use tools, and correct their own mistakes. The guidance focuses on reliable execution, clear operating boundaries, and managing conversation context as tasks grow longer.

• Designing an agent's planning and execution loop
• Breaking a complex goal into manageable tasks
• Reviewing self-correction and reliability patterns

Give an agent rules for consulting long-term memory before work and saving decisions, corrections, and failures afterwards. It helps carry preferences and project conventions across sessions while preserving dated history and making conflicting memories visible.

• Carry preferences and project conventions across sessions
• Record user corrections for future tasks
• Preserve the history of changed decisions

Review and rewrite architecture decision records, documents that explain design choices, from a newcomer’s perspective. Identify missing context, buried decisions, and temporary details, then reorganize each document so future readers can follow the reasoning.

• Reviewing design decisions from a new team member’s perspective
• Making an ADR’s decision and reasoning easier to follow
• Removing temporary project context from long-lived records

Consult another agent for specialist advice, reviews, or workspace context through QwenPaw commands. Keep follow-up conversations connected using session identifiers, and submit longer tasks in the background while continuing other work and checking results later.

• Ask a specialist agent for advice
• Request a second opinion or review
• Consult an agent with relevant files or context

Use built-in tools to ask another agent for advice, a plan, or a review. Choose from available agents, preserve context across follow-up conversations, and send longer tasks to the background with status checks and configurable time limits.

• Ask another agent to propose a plan
• Request a review or help with a decision
• Continue a consultation with its previous context

Coordinate with other QwenPaw agents when a task needs their expertise, files, or context. Discover available agents, exchange messages, continue existing conversations, and submit longer tasks in the background while other work proceeds.

• Ask a specialist agent for a second opinion
• Consult another agent about its files or context
• Delegate analysis or reports as background tasks

Consult another AI agent for advice, proposed solutions, reviews, or decision support. Choose from available agents, preserve context across follow-up conversations, and submit longer tasks in the background with status checks.

• Ask another agent for a second opinion
• Request a review of a proposed solution
• Continue a conversation with its previous context

Guides the use of RWKV, a language model architecture that carries information forward in a compact state. Covers streaming text generation, long document processing, fine-tuning, and memory comparisons with Transformers.

• Build streaming text generation
• Process long documents in chunks
• Compare memory needs with Transformers

Build and evaluate Mamba models for long sequences using state-space models, which summarize past inputs in an internal state. Covers setup, pretrained text generation, differences between Mamba-1 and Mamba-2, and speed comparisons with Transformers.

• Generate text with pretrained Mamba models
• Compare Mamba-1 and Mamba-2 configurations
• Benchmark generation speed against Transformers

Write Git commit messages that preserve why changes were made. The skill keeps Conventional Commits subject lines and adds structured body entries for user intent, decisions, rejected alternatives, constraints, and lessons supported by the session context.

• Recording the purpose behind a feature
• Preserving design decisions and rejected alternatives
• Documenting constraints and implementation lessons