Inspects database health by analyzing slow queries, identifying bottlenecks, and checking Postgres performance indicators. Use when diagnosing database issues, checking slow queries, inspecting DB health, finding performance bottlenecks, or troubleshooting Supabase Postgres problems.

Configures pgvector HNSW indexes with correct distance operators. Enforces inner product operator for normalized embeddings instead of cosine distance. Use when setting up vector search, creating embedding indexes, configuring semantic search, or choosing between HNSW and IVFFlat index types.

Generates code verifier/challenge pairs, exchanges authorization codes for sessions, sets secure HTTP-only session cookies, and configures server-side callback handling to implement the Proof Key for Code Exchange auth flow. Replaces implicit token-in-fragment flows by removing hash-based token extraction and localStorage usage. Use when implementing Supabase authentication, OAuth login, secure browser auth, SSR auth setup, or modern auth flow migration.

Build, customize, and deploy Ghost CMS themes. Use this skill whenever the user mentions Ghost themes, Ghost CMS, Handlebars templates (.hbs files), Ghost Admin, Ghost membership/subscription integration, Ghost custom settings, or Ghost content API — even if they don't say "theme" explicitly. Trigger on: building a blog theme, creating a Ghost site, editing .hbs templates, adding member-only content, Ghost hero sections, Ghost routing (routes.yaml), Ghost image optimization, Ghost dark mode, Ghost search, Ghost deploy, gscan validation, Ghost JSON-LD/SEO, or any mention of {{ghost_head}}, {{ghost_foot}}, {{#foreach}}, {{#get}}, {{img_url}}, {{asset}}, `@custom`, `@member`, or Portal. Also use when the user has an existing Ghost theme they want to modify, extend, or debug — not just for new themes.

Use when building joi schemas, validating input data, defining custom types, conditional validation with .when(), cross-field references, custom error messages, or writing joi extensions. Standalone package that integrates with the hapijs ecosystem.

Diagnoses and optimizes SQL Server database performance. Use when diagnosing slow T-SQL queries, tuning indexes, reading execution plans, fixing parameter sniffing, optimizing batch operations, reducing transaction log bloat, troubleshooting locking and blocking, configuring tempdb, or when a query that used to be fast is now slow. Also use when writing high-throughput INSERT/UPDATE/DELETE operations, implementing minimal logging, designing covering indexes, or analyzing wait statistics.

Use when preparing data for machine learning from SQL Server — feature engineering in T-SQL, building training/test datasets, statistical aggregations for ML pipelines, sampling strategies, data normalization and encoding in SQL, writing queries that feed pandas or scikit-learn, exporting to Parquet or CSV for model training, or when a data scientist asks for a 'feature table' or 'training set' from a SQL Server database.

Use when writing T-SQL for business intelligence, analytics, or reporting. Includes building summary reports with GROUPING SETS, ROLLUP, and CUBE, writing time-series queries with date bucketing, creating pivot/unpivot transformations, generating tally/numbers tables for gap-filling, building running totals and moving averages with window functions, writing year-over-year comparisons, designing materialized views for dashboards, or producing CSV/JSON exports from SQL Server.

Writes, optimizes, and debugs T-SQL queries. Explains SQL Server internals, troubleshoots performance issues, and guides database administration tasks including backup/restore, high availability, security, and index design. Use when the user asks about T-SQL syntax, SQL Server administration, query performance, stored procedures, indexes, locking, transactions, backup/restore, high availability, security, or any MSSQL-related topic — even without saying 'SQL Server' explicitly. Also trigger on terms like SSMS, tempdb, bcp, sqlcmd, MSSQL, sp_executesql, NOLOCK, columnstore, Hekaton, RCSI, param sniffing, or execution plan.

Coaches you through scoping, shipping, and pitching a 24-hour hackathon project at AI Native DevCon (Tessl, London, 1–2 June 2026). Spec-first, track-aware, demo-obsessed. Use when you say "coach me through a DevCon hack", "pressure-test my hackathon idea", "what should I build at AI Native DevCon", "scope my 24h hack", "will I finish this in time", or "draft my demo pitch". Refuses to let you write code before a one-page spec exists.

Implements and enforces full DDD (domain-driven design) conventions for TypeScript/Next.js projects using functional patterns. Validates folder structure against DDD layers, enforces dependency rules between domain/application/infrastructure layers, generates domain module scaffolding, and reviews code for bounded context violations. Use when the user asks about domain-driven design, project architecture, bounded contexts, layer separation, clean architecture, hexagonal architecture, ports and adapters, or organising a TypeScript/Next.js codebase with a domain model, aggregate, value object, or ubiquitous language.

Rigorous thirteen-part synthesis of a text or talk: deep summary, insights, structure, critique, framework rebuild, and CEO-level takeaways. Triggers: reading synthesis, synthesize this, deep dive, rigorous analysis, deconstruct, book analysis, article analysis, essay breakdown, intellectual synthesis, multi-dimensional analysis, executive summary of ideas, framework extraction. Uses: Read (and related file tools) for attached sources; WebSearch or WebFetch when comparands are missing or context is thin. Outputs: single structured markdown message with fixed section headers per rules/output-sections.md.

Remove AI-writing tells and rewrite prose to sound natural and human. Triggers: humanise, humanize, de-AI, deAI, sound natural, less robotic, remove AI tells, AI writing, chatbot tone, Wikipedia AI writing signs, voice match, voice calibration. Uses AskUserQuestion for intake; Read/Write for files; WebSearch/WebFetch when the user wants terminology or usage research; Bash when applying edits in a repo. Outputs: interactive Q&A then rewritten text (and optional brief edit log). Do NOT use for: inventing facts, changing the author's stance, weakening safety or compliance language without explicit instruction, or skipping intake when the user's goals and preservation rules are unclear.

List pending skills and rules on the NAS staging area. Shows what the agent has created or updated that hasn't been promoted to tiles yet. Use before running promote, or when the user asks what's on staging.

Kill a running agent container on the NAS by Telegram group JID. The orchestrator respawns a fresh container on the next message. Does NOT delete registration or group folder. Use when a container is stuck, stale, or needs a fresh start.

Promote agent-created skills and rules from NAS staging to tile GitHub repos via a full PR lifecycle — opens a PR, summons Copilot, iterates fixups until the review is clean, then merges so GHA publishes. Use when there are new items on staging, after check-staging shows pending items, or when asked to deploy skills, push to production, or publish rules to a tile repo.

Lists permitted and prohibited actions, blocks disallowed content types, and responds to permission queries in shared or public group settings. Use when joining a new group, when unsure about rules, permissions, or boundaries, when someone asks what you are allowed to do here, or when operating in a public channel or untrusted group chat environment.

Turn rough ideas into clear work orders before planning or building. Use when request is vague like "add a button", "make it better", "fix the thing". Triggers on ambiguous or underspecified requests. Produces a brief with scope, acceptance criteria, and stop conditions.

Launch Chrome with an unpacked extension and test its UI via CDP. Auto-installs Chrome for Testing if needed. Loads the extension, opens sidepanel/popup/options page, and hands off to cdp-connect for interaction (click, type, screenshot, ax-tree). Handles Chrome 137+ branded build restrictions (Extensions.loadUnpacked via pipe), sidepanel user gesture requirements, and React input quirks. Use when you need to test a Chrome extension's UI, automate extension interactions, or validate extension behavior on a target page. Triggers on: chrome extension test, test extension, load unpacked extension, extension sidepanel, extension popup, test chrome extension, extension testing, chrome extension automation, ext pilot, cdp extension.

Search for Israel Rail train schedules using the railil CLI. Use when the user asks about Israeli trains, Israel Rail schedules, train routes in Israel, or mentions the railil tool. Find routes between stations with fuzzy search, filter by date/time, and output in various formats (JSON, Markdown, Table).

Prepare any webpage for clean interaction by detecting and removing disruptive overlays (cookie banners, GDPR consent, modals, popups, newsletter signups, paywalls, login walls). Uses a cached database of 300+ known CMPs (Consent-O-Matic + EasyList) combined with heuristic DOM scanning. Produces portable JS recipes for any browser tool (Playwright, CDP, cmux-browser). ALWAYS use this skill before taking screenshots, scraping content, or automating interaction on any webpage that might have overlays blocking the view or preventing interaction. Triggers on: page prep, clean page, remove overlays, dismiss cookie banner, page blocked, overlay cleanup, consent banner, prepare page, unblock page, clear popups, cookie popup.

Review Claude Code auto memory accumulated during a task and promote valuable findings to shared project config (CLAUDE.md, .claude/rules/, or global CLAUDE.md). Use this skill when finishing a branch, completing a task, wrapping up work, or anytime you want to review what Claude learned and decide what the team should know. Also use when the Stop hook reminds you about pending memory entries. Triggers on "triage memory", "review memory", "promote memory", "memory triage", "what did we learn", "clean up memory", "share learnings", "anything worth keeping", or when the user asks about knowledge accumulated during a session.