nsharandroidnstudio
- 3 skills
- 0 followers
- 17 hours ago last updated
- ▌ Reduce LLM Token Costs · nsharandroidnstudioUse when an LLM agent, RAG pipeline, or multi-step tool workflow is burning too many tokens/dollars per request — cut token cost ~90% by compressing tool defs, summarizing telemetry/logs, compressing RAG context, replacing conversation history with external state, hierarchical compression, small-model preprocessing, prompt caching, and conversation chaining. Reach for it when a single request exceeds a few thousand tokens of "plumbing" (system prompt + tools + history + retrieved chunks + raw data) rather than useful signal.
- ▌ Deep Agents Architecture · nsharandroidnstudioUse when an agent must run long-horizon, open-ended, parallelizable tasks (deep research, large multi-file code changes, hundreds of record lookups) and a single-loop agent drifts after 5-20 steps. Encodes the four-pillar deep-agents pattern — planning tool (write_todos), virtual filesystem for context offload, isolated subagents, long-term memory — plus the token/KV-cache economics that decide whether it's viable and a step-count rule for when NOT to use it. Reach for it past ~15-20 steps with separable sub-tasks; stay single-loop below that.
- ▌ Writing Tools For Agents · nsharandroidnstudioUse when designing, building, or refactoring tools (function calls / MCP tools) that an AI agent will invoke — how to consolidate tools around workflows instead of wrapping API endpoints, namespace them, return high-signal natural-language context instead of raw IDs/UUIDs, offer concise vs detailed response formats, stay token-efficient with pagination/filtering/truncation, write actionable error messages, prompt-engineer tool descriptions, and drive the whole thing with evaluation-driven development. Reach for it whenever an agent misuses tools, calls too many, chokes on raw tool output, or you're standing up a new MCP server.