codejunkie99

Use when the user wants to craft a sincere, accountable apology in both written and spoken versions, without deflection. Triggers include "draft an apology", "how should I apologize", "say sorry to".

Use when the user wants to prepare for a hard conversation with empathetic-but-direct opening, defensiveness handling, and resolution paths. Triggers include "difficult conversation", "hard talk", "tough conversation prep".

Use when the user wants three energy-level variants of an elevator pitch (bold, conversational, data-driven). Triggers include "elevator pitch", "pitch this product", "30 second pitch".

Use when the user wants to draft spoken feedback using observation/impact/expectation/support framework. Triggers include "give feedback", "feedback to my report", "performance feedback".

Use when the user wants a tight presentation outline with hook, problem, solution, evidence, objection handling, and clear CTA. Triggers include "outline a presentation", "talk outline", "structure my pitch deck".

Use when the user wants plain-English interpretation of a dataset including patterns, surprises, and misleading aspects. Triggers include "analyze this data", "interpret these numbers", "what does this data mean".

Use when the user has multiple research sources on a topic and wants genuine synthesis of themes, contradictions, and gaps — not separate summaries. Triggers include "synthesize this research", "compare these sources", "research synthesis".

Use when the user wants to analyze survey results for findings, consensus, division, surprises, and actionable recommendations. Triggers include "analyze survey", "survey results", "what does this survey show".

GateFlow terminal console inspired by OpenClaw local TUI workflows. Shows workspace status, component inventory, tool health, map readiness, release readiness, and command shortcuts from one terminal surface.

GateFlow release readiness workflow. Validates plugin manifests, marketplace metadata, docs index coverage, root mirrors, release notes, and component counts before a version tag is created. Use when preparing, checking, or cutting a GateFlow plugin release.

FuseSoC build system integration for GateFlow. Generates .core files and drives synthesis/simulation through Edalize backends (Vivado, Quartus, open-source tools). Use when the user needs to create a FuseSoC core file, build with Edalize, or integrate RTL…

SystemVerilog learning mode — generates exercises, reviews solutions, and teaches RTL design patterns. Use when the user wants to learn SystemVerilog, practice hardware design, get exercises, or understand verification methodology.

Figures out what kind of digital hardware design task the user wants to do, then hands off to the right specialist. Use when the request is unclear, multi-step, or needs help deciding whether to simulate, synthesize, lint, or implement.

Primary SystemVerilog/RTL orchestrator for GateFlow. Routes to specialist agents, runs verification, and iterates until working. Use when the user wants to create, test, fix, or implement any RTL design — FIFO, UART, AXI, state machines, or any digital…

Summarize Verilator, lint, or simulation output into a readable, actionable format. Use when the user wants to understand build output, lint errors, or simulation results from a Verilator or EDA tool run.

Codebase architect - Maps and documents SystemVerilog projects. This skill should be used when the user wants to understand a codebase structure, generate architecture documentation, or onboard to a new RTL project. Example requests: "map this codebase",…

Search and attach agent or subagent context across Claude Code, Codex, OpenCode, and Cursor when the user types @ or a tool mention.

Recall and store durable cross-tool memory through the external Brain CLI.

Export approved, redacted agent runs into local retrieval artifacts and evaluation cases.

Inspect cross-tool agent activity, usage, scheduled work, and local reports.

Reproduce, isolate, and verify the root cause of a software failure.

Check release readiness and request human approval before deploying.

Apply the project DESIGN.md visual design contract and its existing tokens.

Perform Git operations with branch protections and verification checks.

Build or operate a company-specific harness using bounded work packets, evidence-led artifacts, model-route records, UI artifacts, and dashboards.

Package company work as a reviewable artifact with provenance, evidence, authority, unresolved questions, and a recorded review state.

Deliver a scoped code or app change with repository evidence, focused verification, reviewable artifacts, and explicit release authority.

Build an inspectable operational dashboard with metric definitions, data lineage, freshness, permissions, states, and verification evidence.

Turn a company work packet into a usable UI with explicit users, tasks, data contracts, states, implementation details, and verification evidence.

Convert a company request into a bounded packet with explicit outcome, scope, authority, inputs, outputs, and checks.

Record an eligible model tier, rationale, and fallback for a bounded company work packet.

Produce a source-led customer research artifact that separates evidence, claims, inference, uncertainty, and next actions.

Write a short launch caption, quote-post hook, or agent handoff for an existing article. Use for concise copy that communicates the payoff without rewriting the full article.

Find article ideas at the intersection of audience questions, current discussion, and the author's actual experience. Use for topic research, angle selection, or developing a thesis from notes and sources.

Package an article and its supporting prompts, diagrams, or project files into a documented, reusable folder or GitHub repository. Use when readers need downloadable material or an agent-readable handoff.

Review an article, title, thumbnail, and supporting files for factual support, reader usefulness, and consistency. Use for a publication-readiness check or a focused critique of a draft.

Turn an author's experience into an original, useful X article with a coordinated title and thumbnail. Use for the full workflow from an idea or voice dump to a reviewed draft and reusable files.

Design article thumbnail concepts, generation prompts, and precise image revisions. Use for reference selection, readable workflow diagrams, visual style extraction, or coordinated title and thumbnail packaging.

Write or refine X article titles, opening promises, and accessible positioning. Use when choosing a reader outcome, comparing titles, or aligning the title with an outline and thumbnail.

Draft, condense, or format an original X article from the author's notes, voice transcripts, and evidence. Use for full articles, early TLDRs, reproducible explanations, or numbered and arrow-list formatting.

Implement a task with DeepSeek V4 Pro and verify it. Use when the user invokes /build or wants the implement utility.

Run the task on DeepSeek V4 Pro. Use when the user invokes /deepseek or wants deeper DeepSeek reasoning.

Run the task on DeepSeek V4 Flash. Use when the user invokes /flash or wants the fast DeepSeek utility.

Run the self-improving DeepSeek/Luna/Sol loop. Use when the user says "/loop", "$loop start", or asks to start a self-improving loop with a task, including a repeating 10-minute cadence.

Use Luna to watch, grade, or run bounded worker tasks. Use when the user invokes /luna or wants the Luna utility.

Use Sol to review work or rewrite prompts. Use when the user invokes /sol or wants the Sol utility.

Show the utility-to-model menu for slash commands. Use when the user invokes /utility or asks for the model menu.

This skill should be used when the user is working on any SystemVerilog, Verilog, HDL, RTL, FPGA, or ASIC development task. Provides behavioral rules for GateFlow: auto-chain workflows, context-aware actions, and seamless tool execution without requiring…

Provides RTL design patterns, synthesis guidelines, and coding templates. Use when the user mentions 'module design', 'FSM', 'state machine', 'pipeline', 'synthesizable', 'parameterize', 'parameter', 'localparam', 'clock domain', 'CDC', 'FIFO', 'register…

This skill should be used when the user mentions 'lint error', 'width mismatch', 'inferred latch', 'undriven signal', 'fix warnings', 'Verilator error', 'WIDTHTRUNC', 'BLKSEQ', or asks to 'check my code' or 'find errors'. Provides error-specific fix patterns…

This skill should be used when the user asks to 'write a testbench', 'test this module', 'add assertions', 'create test stimulus', 'verify the design', 'add clock generation', or 'write a self-checking test'. Provides testbench templates, clock/reset…

Provides verification methodology, SVA assertion patterns, coverage techniques, and testbench templates. Use when the user mentions 'testbench', 'assertion', 'SVA', 'property', 'sequence', 'coverage', 'covergroup', 'coverpoint', 'bins', 'cross coverage',…

This skill should be used when the user mentions 'waveform', 'VCD file', 'simulation results', 'signal trace', 'debug the output', 'what happened in sim', 'show me the waves', or 'analyze timing'. Provides VCD parsing, clock detection, anomaly identification,…

This skill should be used when the user asks to 'design a module', 'write SystemVerilog', 'implement an FSM', 'create a pipeline', or works with .sv/.svh/.v/.vh files. Provides modern SV conventions, coding patterns, and synthesizability guidance.

Map reference images into selectable named elements, produce evidence-labeled reconstruction JSON, and compile it into prompts for reconstruction, remixing, and scoped edits. Use when a person wants to understand, recreate, or selectively change an image.

Analyze image references into named visual elements and reconstruction JSON, then compile prompts for recreating, remixing, or selectively editing them. Use when the user wants to understand a reference, reuse its visual decisions, or preserve specific parts…

Generate or edit images with a bounded create-review-repair loop, an independent vision reviewer, explicit preservation checks, and saved evidence. Use for image work that needs iterative checking, or invoke image-loop with a brief or reference image.

Turn multiple inspiration images, links, and notes into named visual ingredients, diverse combinations, and generated images. Use for moodboards, reference mixing, or creative exploration with one batch or a bounded human- or vision-judged iteration loop.

Guide image edits through questions, numbered visual maps, named elements, and preservation checks. Reverse-engineer reference images into structured JSON for typography, palette, composition, inferred layers, grading, and image types. Use for image editing,…

Extract a supplied image into structured JSON with typography, palette, grading, composition, named elements, inferred layers, lighting, and evidence confidence.

Build and maintain a design knowledge system from website URLs and screenshots. Captures inspiration, extracts reusable patterns, synthesizes connections, and builds persistent design memory. Ripgrep-first — no external LLM required.

Conversational onboarding that generates a personalized design knowledge system. Asks about your design practice, then creates identity, methodology, MOCs, and processing pipeline.

Build and query a relational markdown design memory using design-brain-memory and Agent Browser CLI.

When to delegate to `cavecrew-investigator` (locate code), `cavecrew-builder` (1-2 file edit) or `cavecrew-reviewer` (diff review) instead of working inline or using `Explore`. Their output is compressed, so main context lasts longer.

Ultra-compressed communication mode that cuts output tokens while keeping technical accuracy. Levels: lite, full, ultra and the wenyan variants. Use for /caveman, "caveman mode", "talk like caveman", "be brief" or "less tokens".

Use the locally authenticated Claude Code Fable 5.1 model only as the orchestrator for a task, then execute implementation with GPT-5.6 Luna or DeepSeek V4 Flash. Use when the user invokes $fable or asks Fable to orchestrate Codex agents.

Teaches an agent graph engineering — both halves. Knowledge graphs (ontology design, entity/relation/event extraction, fusion, GraphRAG/memory serving; distilled and translated from Southeast University's graduate Knowledge Graph course,…