Enable long-running background task support with LongRunningAgentServer. Use when: (1) Agent tasks may exceed HTTP timeout (~120s), (2) User wants background/async execution, (3) User says 'long running', 'background tasks', or 'async agent'.

Modify agent code, add tools, or change configuration. Use when: (1) User says 'modify agent', 'add tool', 'change model', or 'edit agent.py', (2) Adding MCP servers to agent, (3) Changing agent instructions, (4) Understanding SDK patterns.

Modify agent code, add tools, or change configuration. Use when: (1) User says 'modify agent', 'add tool', 'change model', or 'edit agent.py', (2) Adding MCP servers to agent, (3) Changing agent instructions, (4) Understanding SDK patterns.

Modify agent code, add tools, or change configuration. Use when: (1) User says 'modify agent', 'add tool', 'change model', or 'edit agent.py', (2) Adding MCP servers to agent, (3) Changing agent instructions, (4) Understanding SDK patterns.

Modify agent code, add tools, or change configuration. Use when: (1) User says 'modify agent', 'add tool', 'change model', or 'edit agent.py', (2) Adding MCP servers to agent, (3) Changing agent instructions, (4) Understanding SDK patterns.

Modify agent code, add tools, or change configuration. Use when: (1) User says 'modify agent', 'add tool', 'change model', or 'edit agent.py', (2) Adding MCP servers to agent, (3) Changing agent instructions, (4) Understanding SDK patterns.

Run and test the TypeScript LangChain agent locally. Use when: (1) User wants to test locally, (2) User says 'run locally', 'test agent', 'start server', or 'dev mode', (3) Debugging issues.

Reference for Packmind CLI listing commands. This skill should be used when an agent needs to discover available standards, commands, or skills in the Packmind organization.

Capture new learnings discovered during code exploration or user interaction that were not present in context, memories, or codebase documentation. This skill should be invoked when Claude discovers something new about the codebase, tools, workflows, or domain knowledge that could benefit future sessions.

Internal maintainer SOP for version bumps, release PRs, tagging, publishing, and post-publish verification in this repository.

Publish a per-version release report for `eval/token_ceiling_formula.py` every time `FORMULA_VERSION` bumps — the exact configuration (formula, signal weights, calibration constants), a ranked gap list for the next contributor, and what's settled and not worth re-relitigating. Why this is its own skill rather than folding into `model-right-sizer-research-report`: every claim must interweave WHY it matters, in the same breath as WHAT changed — it stays decision-support only if a reader can tell, without a second file, what breaks (wasted spend, false alarms, undetected overruns, a re-biased fleet of budgets) if a number or gap is wrong. Never a new-finding surface — synthesizes only from committed results files. Also the tool for BACKFILLING a report for a past version that shipped before this skill existed. Use when someone says "write the release report for this version", "version the token ceiling formula", "backfill a release report for v0.x", or after any `FORMULA_VERSION` bump.

Tune model-right-sizer's wording knobs (`eval/tuning/knobs.py`) against a REAL, already-measured build's actual token spend — not the synthetic benchmark `model-right-sizer-prompt-tuning` searches, nor a fresh build per candidate. Picks a task from `overfitting_guard.py`'s `HOLDOUT_TASKS` registry (real actual/budgeted pairs already recorded), dispatches 3 INDEPENDENT BLIND dry-runs per candidate (no calibration-ledger access), averages the budget across draws (single-draw noise can flip within/over-budget calls), maps to real actuals, scores via `classify_budget_adherence` + `score_candidate`, diagnoses the miss pattern, proposes ONE wording change, and re-runs to check improvement. n stays fixed per task — flags rather than silently continues once squeezing looks like overfitting. Use when someone says "tune the knobs against this blueprint/build", "iterate the dry run with no prior context against the real actuals", "keep tuning until N%", or "how close does a blind estimate get to the actual cost".

Advanced 7-step hierarchical design prompt generator for AI web development tools (Lovable, Cursor, Bolt). Generates domain-aware, user-journey-based design prompts with emotional design considerations. Triggers on "디자인 프롬프트", "웹 디자인", "Lovable 프롬프트", "랜딩페이지 만들어줘", or any AI web builder prompt requests.

Use when learning Rust concepts. Keywords: mental model, how to think about ownership, understanding borrow checker, visualizing memory layout, analogy, misconception, explaining ownership, why does Rust, help me understand, confused about, learning Rust, explain like I'm, ELI5, intuition for, coming from Java, coming from Python, 心智模型, 如何理解所有权, 学习 Rust, Rust 入门, 为什么 Rust

Use when designing resource lifecycles. Keywords: RAII, Drop, resource lifecycle, connection pool, lazy initialization, connection pool design, resource cleanup patterns, cleanup, scope, OnceCell, Lazy, once_cell, OnceLock, transaction, session management, when is Drop called, cleanup on error, guard pattern, scope guard, 资源生命周期, 连接池, 惰性初始化, 资源清理, RAII 模式

CRITICAL: Use for error handling. Triggers: Result, Option, Error, ?, unwrap, expect, panic, anyhow, thiserror, when to panic vs return Result, custom error, error propagation, 错误处理, Result 用法, 什么时候用 panic

CRITICAL: Use for mutability issues. Triggers: E0596, E0499, E0502, cannot borrow as mutable, already borrowed as immutable, mut, &mut, interior mutability, Cell, RefCell, Mutex, RwLock, 可变性, 内部可变性, 借用冲突

The Algolia side of a Vercel Marketplace integration: which Algolia variables it injects, what the Algolia CLI and the browser each need them to be named, and how to reach a validated first search. Use when the user is on Vercel and wants Algolia added from the Marketplace — "set up Algolia on Vercel", "vercel integration add algolia" — or when Marketplace-injected ALGOLIA_APP_ID / ALGOLIA_SEARCH_API_KEY / ALGOLIA_WRITE_API_KEY need to be wired into an app and proven with a query. Vercel's own mechanics — teams, plans, terms, billing, environment commands — belong to the Vercel docs, not here. Do NOT use merely because a project is Next.js — Next.js alone does not mean Vercel. Do NOT use when the user has a standalone Algolia application they intend to keep. Do NOT use for records, settings, synonyms, rules or keys on an existing application (algolia-cli), for search UI (instantsearch), for read-only search and analytics (algolia-mcp), or for accounts created outside Vercel (algolia-quickstart).

Design and build launch landing pages that convert visitors into users. Use this skill whenever someone asks to create a landing page for a product launch, build a feature page, design a signup page, or structure a product marketing page. Also trigger for "make a page for this launch," "I need a landing page," "help me design the product page," or any request to create a web page tied to a product release or feature announcement. Covers page structure, hero design, SEO integration, developer-first patterns, urgency mechanics, waitlist flows, and conversion optimization.

エージェントのCLI/モデルをライブ切替するスキル。settings.yaml更新→/exit→新CLI起動→ pane metadata更新を一発で実行。Thinking有無も制御。 「モデル切替」「Sonnetにして」「Opusに変えて」「足軽全員切替」「Thinking切って」で起動。