Research & Data

Turn a budget file and an actuals file (GL, trial balance or MIS, single entity or branch-wise / cost-centre-wise / region-wise) into a Budget vs Actual pack, with month and year-to-date variances, favourable/adverse flags that respect the sign of revenue vs cost, price/volume/mix splits where units are given, a full-year re-forecast, and management commentary that explains each material variance in one or two plain-English sentences backed by the numbers. Ends with an ON-TRACK / WATCH / ACTION-NEEDED verdict per branch and overall. Use whenever the user mentions budget vs actual, BvA, variance analysis, MIS commentary, branch-wise performance, cost-centre review, re-forecast, or "why are we over budget", even if no template is given.

Plan a rigorous topic-specific web investigation, compare independent evidence and create a cited research report with verified source accounting.

Draft, review, and schedule distinct social posts with the TimeToPost MCP. Use when asked to plan a social calendar, draft a post or thread, check approval drafts, choose a posting time, or schedule content.

Captures, sanitizes, and maintains high-value cross-session state in persistent memory files (MEMORY.md) while strictly protecting sensitive data and discarding ephemeral noise.

Content Research command - find high-demand, low-supply content gaps. Use only when the user invokes cr-gaps or asks for the content-research "gaps" action. Entry point for the content-research skill.

Content Research command - analyze competing creators and their content. Use only when the user invokes cr-competitors or asks for the content-research "competitors" action. Entry point for the content-research skill.

Content Research command - run the complete content research workflow. Use only when the user invokes cr-full or asks for the content-research "full" action. Entry point for the content-research skill.

Use when caching GraphQL reads in a NestJS (code-first) app — wiring a stale-while-revalidate (SWR) cache module once, deciding what to cache and where to mount it (service, field, operation, or CDN), and keeping it leak-safe. Triggers when you need SWR, single-flight dedupe, tag invalidation, per-user vs public scoping, cache observability, a stampede or blocking-miss at TTL expiry, a shared-key data leak, a dev/ops switch that disables caching entirely, or a one-time setup reference (module, file placement, storage, referencesTTL/GC, caller identity, decorator runtime, fail-open).

Search the web before answering or carrying out user requests, then use checked sources and citations. Use when the user wants web-first answers or research before each task, especially with models that need external knowledge.

Reviews macOS Swift 6+ code for modern idioms, SOLID principles, SwiftData patterns, and concurrency best practices. Use when reviewing macOS code quality or asking about best practices.

为市场调查项目选择和实施与研究问题、变量类型、抽样设计匹配的定量、质性及文本分析，检查测量、模型、稳健性和解释边界。用于市调分析预案、统计方法选择、结果解释和方法纠错。

Change the shape of stored data (add, rename or remove a field on Link) safely, including a versioned upgrade of existing data files. Use when a task needs a new or changed stored field.

Run a human-in-the-loop development loop for any product - Claude acts as lead agent (understands the request, researches, splits work, reviews, integrates) and dispatches implementation to a coding agent (Codex CLI by default) running in isolated git worktrees, with scheduler-enforced verification, feedback delivery and integration gates, while the user tests the real product and decides. Use when the user wants Claude to "let Codex do it", orchestrate coding agents on a codebase, set up an agent development workflow, run several coding tasks in parallel worktrees, or keep iterating on user feedback (screenshots, "this looks wrong", "there's a bug").

Make an informed non-trivial decision via a three-step interrogation → criteria → research-handoff protocol. Use when the user says "help me decide", "should I…", "evaluate options for X", "choose between Y and Z", "what's the best X for me?", "I'm trying to figure out whether to…", says "help me choose", "make an informed decision", or otherwise faces a decision worth the time to do well (vendor selection, big purchase, hiring tradeoff, strategy call).

Traces the full call chain from a given symbol or code location up to all entry points that can reach it (HTTP controllers, main functions, page components, routes, etc.). Works for any language indexed by CodeGraph. Use for blast-radius analysis before modifying a function, method, or class.

Comprehensive quality verification: spec compliance, lint, type-check, tests, cross-layer data flow, code reuse, and consistency checks. Use when code is written and needs quality verification, before committing changes, or to catch context drift during long sessions.

Guides collaborative requirements discovery before implementation. Creates task directory, seeds PRD, asks high-value questions one at a time, researches technical choices, and converges on MVP scope. Use when requirements are unclear, there are multiple valid approaches, or the user describes a new feature or complex task.

Deep bug analysis to break the fix-forget-repeat cycle. Analyzes root cause category, why fixes failed, prevention mechanisms, and captures knowledge into specs. Use after fixing a bug to prevent the same class of bugs.

Build an onboarding map of an unfamiliar repository covering purpose, entry points, how to run it, main modules, data flow and risky areas, with file references. Use when asked to explain, map or onboard onto a codebase.

Pull the people who liked or commented on any LinkedIn post and segment them by ICP fit (peer / aspirational / prospect / other). Produces an engager roster, tier breakdown, and outbound action lists (follow back, comment-drop, DM-able with one-line openers). Powered by Apify, no LinkedIn login. Triggers on \"who liked my post\", \"who engaged\", \"engagers report\", \"audience analytics\". Not for tracking author replies to your comments (use linkedin-thread-monitor).

Download open-access biomedical article PDFs from PMC with offline Proof-of-Work solving. Triggers: 下载文献PDF, PMC PDF, 403/404 on PDF link, Preparing to download, reCAPTCHA page, OA full text, supplementary files, download paper PDF.

Add cached official-asset and product research preflight before reference selection and build authorization.

>- Run Codex CLI (exec, resume) for parallel work, delegating tasks to another model, or background code generation. Handles code analysis, refactoring, and automated editing. Includes Spark vs standard guidance. Triggers: "codex", "run codex", "use codex to".

Use when the user asks to run Codex CLI (codex exec, codex resume) or references OpenAI Codex for code analysis, refactoring, or automated editing