Relentless interview mode that drives toward shared understanding of a plan or design. Walks every branch of the decision tree one question at a time, each with a recommended answer. Use when user says "grill me", "interview me", or invokes /grill-me.
Executes saved browser automation workflows with consistent setup, teardown, and reporting. Loads workflow files and runs them through the browser skill. Use for repeatable browser automations like scraping, form-filling, or monitoring.
Ultra-compressed communication mode. Cuts token usage ~75% by dropping filler, articles, and pleasantries while keeping full technical accuracy. Use when user says "caveman mode", "talk like caveman", "use caveman", "less tokens", "be brief", or invokes /caveman.
Disciplined diagnosis loop for hard bugs and performance regressions. Reproduce, minimise, hypothesise, instrument, fix, regression-test. Use when user says "diagnose this", "debug this", reports a bug, says something is broken / throwing / failing, or describes a performance regression. Triggers: "diagnose", "debug this", "why is X broken", "X is throwing", "X is failing", "perf regression", "this got slow", "intermittent failure", "flaky test".
Text-to-speech, sound effects, music generation, and audio processing using the ElevenLabs API via the el CLI. Use for voice generation, audio content creation, sound design, or audio processing.
Generate or edit images using AI image models. Currently supports Google Gemini (Nano Banana Pro for creation, Nano Banana for editing). Provider-agnostic design. Use for image generation, visual content creation, or image editing tasks.
Grilling interview that stress-tests a plan against the project's existing domain language and documented decisions, sharpening terminology and updating CONTEXT.md and ADRs inline as decisions crystallise. Use when user says "grill me with docs", "grill on the design", "challenge this against our context", or invokes /grill-with-docs.
| name | grill-me |
| description | Relentless interview mode that drives toward shared understanding of a plan or design. Walks every branch of the decision tree one question at a time, each with a recommended answer. Use when user says "grill me", "interview me", or invokes /grill-me. |
Interview the user relentlessly about every aspect of a plan or design until both sides reach a shared understanding. Walk each branch of the decision tree, resolving dependencies between decisions one at a time. Works for any subject — a marketing plan, a doc outline, a product decision, a code change.
Interview Relentlessly Toward Shared Understanding
/grill-me/grill-meWalk Each Branch of the Decision Tree
One Question at a Time, With a Recommended Answer
Resolve from Source When the Question Allows