tekhne

Compares deployed CloudFormation templates with locally synthesized CDK templates to detect drift, validate changes, and ensure consistency before deployment. Use when the user wants to compare CDK output with a deployed stack, check for infrastructure drift, run a pre-deployment validation, audit IAM or security changes, investigate a failing deployment, or perform a 'cdk diff'-style review. Triggered by phrases like 'compare templates', 'check for drift', 'cfn drift', 'stack comparison', 'infrastructure drift detection', 'safe to deploy', or 'what changed in my CDK stack'.

Challenge, push back, play devil's advocate on AI output. Use when: challenge this, are you sure, push back, prove it, what if you're wrong, devil's advocate, stress test, poke holes, second opinion, sanity check, too confident, really?, question this decision. Subcommands: anchor (committed too fast), verify (facts wrong?), framing (wrong problem?), deep (full devil's advocate in separate context).

Expert assistant for chezmoi dotfiles management. Use when the user is managing dotfiles with chezmoi: adding files, creating templates, encrypting secrets, writing run scripts, syncing across machines, or diagnosing why changes aren't applying. Trigger phrases: 'add to chezmoi', 'make a template', 'chezmoi apply', 'encrypt with chezmoi', 'run script on first apply', 'sync dotfiles to new machine', 'chezmoi diff shows unexpected changes', 'source attribute', 'dot_ prefix', 'once_ script'.

Write an original deadpan two-person political-satire interview sketch (a mock-interview bit, not a monologue, panel show, or sitcom scene) using terse questions, absolutely-confident literal-minded answers, repetition, bureaucratic logic, and escalating contradiction, in the tradition of John Clarke and Bryan Dawe's satirical news interviews. Use this when a user asks for a satirical interview, mock-interview sketch, deadpan two-hander, or "Clarke and Dawe style" bit built around a policy, institution, or current-events topic. Do NOT use this for stand-up monologue jokes, sketch comedy with more than two speaking roles, sitcom-style scenes with stage direction and physical comedy, or any request to reproduce an actual historical Clarke and Dawe transcript verbatim — this skill generates new satire, it does not retrieve or reconstruct existing scripts.

Apply Clean Architecture principles to define layer boundaries, identify dependency violations, and structure domain vs infrastructure code. Use when designing service boundaries, separating business logic from infrastructure, evaluating hexagonal/onion/ports-and-adapters architecture, structuring module layout, or resolving dependency inversion and circular dependency issues.

Given a user-provided set of coding standards (either as direct text input or as a link to a file/URL), systematically review the codebase to identify violations, report findings, and optionally remediate them. The model MUST ask the user for the standards — never guess or assume them.

Build authoritative real-time multiplayer servers with Colyseus 0.17+. Use when implementing rooms, schema state sync, client message validation, matchmaking, authentication, reconnection handling, or server-side anti-cheat constraints. Keywords: colyseus, room lifecycle, schema, multiplayer, websocket, matchmaking, onJoin, onLeave, onDrop, allowReconnection.

Generate a new, reusable comedy-writing SKILL.md for a requested comedic tradition — real, generic, or invented — by extracting its transferable comic mechanics (premise construction, escalation pattern, rhythm, rhetorical devices) into concrete, actionable writing rules instead of vague adjectives, and by converting any living performer's persona into technique rather than voice-imitation. Use this when a user asks to create a comedy skill for a style, or a skill "like" a comedian or show, and wants a complete SKILL.md file back. Do NOT use this to write comedic material directly (that is the generated skill's job, not this one's), for non-comedy skill generation, or to produce a skill whose body just says "be funny" or "make it absurd" without concrete mechanics.

Complete Commander.js CLI framework guidance covering command structure, options, arguments, subcommands, action handlers, version management, and TypeScript integration. Use when: building CLI tools, parsing command-line arguments, implementing subcommands, handling options/flags, creating interactive CLIs, or migrating from other CLI frameworks. Keywords: Commander.js, CLI, command-line, arguments, options, flags, subcommands, action handlers, version, help text, TypeScript, yargs, meow, program, parseAsync, opts, args, variadic, required options, default values, custom help, error handling

Regenerates the project's context index (index.yaml) from the YAML frontmatter across every context file, grouping entries by typology (findings/plans/goals/evidence/guides/follow-ups/handovers/merge-requests/tickets/decisions/notes/research), and validates that each file carries the required frontmatter fields. Also computes which active files are ready to work on, from optional typed `blocks`/`blocked-by` dependency links (scripts/context-ready.sh). Use when the index is stale, context files were added, renamed, or removed, a pre-commit gate blocks a commit because a context file is missing frontmatter, or someone asks what's ready/unblocked/blocked across the context backlog. Do not use it to create new context files (use create-context-file instead), to hand-edit the index, or as a substitute for fixing frontmatter at the source.

Audit a project's Claude Code tool stack (MCP servers, plugins, hooks) against the curated recommendations at Context Radar (thoroc.github.io/context-radar): a 90-plus-tool, 20-layer comparison of context-reduction, memory, and code-nav tools for coding agents. Use when asked to review the Claude Code / MCP tool stack, check for redundant or conflicting tools, decide what to install next, or evaluate a new context/memory/code-nav tool against what's already running. DO NOT use for reviewing application dependencies (npm/cargo packages) or general code review - scope is Claude Code tooling only. Triggers: 'context radar', 'what tools should we use', 'audit our tool stack', 'is X redundant with Y', 'should I install this MCP server', 'tool stack review', 'context reduction tools'.

Generates and formats git commit messages following the Conventional Commits specification. Use when the user asks for help writing commit messages, formatting git commits, wants to write a commit msg for staged changes, needs to produce a changelog entry, or mentions conventional commits, semantic versioning, or version bumps. Analyzes diffs or change descriptions, selects the correct type/scope, writes an imperative-mood header under 72 characters, and composes optional body and footer sections including breaking-change notation.

Create the baseline session context from the in-folder of raw project documents, classifying each file by priority, writing a manifest, summarising oversized files and producing a token-budgeted baseline. Use when bootstrapping project context or setting up the ctx snapshot once per project. Triggers include "create context", "bootstrap context", "setup context", "init context".

Creates a structured, date-stamped context file filed by typology (findings, plans, goals, evidence, guides, follow-ups, merge-requests, tickets, decisions, notes, research) with YAML frontmatter. Use when saving a finding, writing a plan, capturing a guide, recording a decision, logging a follow-up, filing a merge-request note, capturing work as progression evidence, plus ticket write-ups.

Generate an original mockumentary presenter-satire segment, documentary narration plus an expert-interview scene, in a sincere-ignorance deadpan style. The presenter is a supremely confident, credential-free character who states garbled misunderstandings as flat fact; a genuine credentialed expert answers with strained, patient sincerity. Trigger this skill for requests to write, draft, punch up, or extend a mockumentary voiceover, deadpan documentary narration, faux-expert interview scene, satirical documentary script segment. Do NOT trigger for stand-up material, self-aware sketch comedy, roast-style insult comedy, hostile interview dialogue, straight factual documentary writing, reproduction of a real broadcaster's actual presenter, character, verbatim lines. This skill only produces new, original material based on the structural technique, not a copy of real broadcast content.

Extract clean markdown content from web pages using Defuddle CLI, removing clutter and navigation to save tokens. Use instead of WebFetch when the user provides a URL to read or analyze, for online documentation, articles, blog posts, or any standard web page.

Stress-test an unwritten idea or decision by spawning independent subagents in genuinely OPPOSING roles (advocate, skeptic, migration/risk) rather than complementary review lenses — grounded in real repo investigation, not assumption, and concluding in a synthesized verdict rather than a list of opinions. Use for "should we do X", "what are the pros and cons", "I need pushback", "collegial review", "second opinion on this decision" — BEFORE a plan exists. NOT for reviewing an already-written plan file (use plan-review) and NOT for refining a vague question through dialogue with the user (use socratic-method).

Select and apply structural design patterns (Strategy, Factory, Adapter, Observer, etc.) to reduce complexity in class design. Use when evaluating whether a pattern solves a concrete problem, refactoring conditional logic, dealing with too many if/else statements or code smells, resolving inheritance problems, or deciding how to structure classes.

Generates production-ready Dockerfiles with multi-stage builds, layer caching, security hardening (non-root users, minimal base images), health checks, and automatic .dockerignore creation. Validates output using devops-skills:dockerfile-validator with iterative error fixing. Use when a user wants to create, generate, write, or build a Dockerfile or docker image, containerize an app, set up a container, or needs a Dockerfile for Node.js, Python, Go, Java, or another language. Also applies when the user says 'containerize my app', 'docker-compose setup', 'container setup', or 'optimize my Docker build'.

Validates, lints, and secures Dockerfiles by running syntax checking, detecting security vulnerabilities, validating layer ordering, checking for hardcoded secrets, verifying base image tags, and analyzing build optimization. Use when validating Dockerfile syntax, checking security best practices, optimizing image builds, auditing container security, or debugging Dockerfile errors. Applies to all Dockerfile variants (Dockerfile, Dockerfile.prod, Dockerfile.dev, etc.).