Skip to content
Open
Show file tree
Hide file tree
Changes from all commits
Commits
File filter

Filter by extension

Filter by extension

Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
2 changes: 1 addition & 1 deletion README.md
Original file line number Diff line number Diff line change
Expand Up @@ -4,7 +4,7 @@

A Claude skill that writes the accurate prompts for any AI tool. Zero tokens or credits wasted. Full context and memory retention. No re-prompting your way to an answer you should have gotten on attempt one.

**Works with:** Claude, ChatGPT, Gemini, o1/o3, MiniMax, Cursor, Claude Code, GitHub Copilot, Windsurf, Bolt, v0, Lovable, Devin, Perplexity, Midjourney, DALL-E, Stable Diffusion, ComfyUI, Sora, Runway, ElevenLabs, Zapier, Make, and any AI tool you throw at it.
**Works with:** Claude, ChatGPT, Gemini, o1/o3, MiniMax, Cursor, Claude Code, Cortex Code, GitHub Copilot, Windsurf, Bolt, v0, Lovable, Devin, Perplexity, Midjourney, DALL-E, Stable Diffusion, ComfyUI, Sora, Runway, ElevenLabs, Zapier, Make, and any AI tool you throw at it.

---

Expand Down
11 changes: 11 additions & 0 deletions SKILL.md
Original file line number Diff line number Diff line change
Expand Up @@ -165,6 +165,17 @@ Identify the tool and route accordingly. Read full templates from [references/te

---

**Cortex Code (Snowflake's CLI coding agent)**
- Agentic like Claude Code — runs tools, edits files, executes SQL, manages Snowflake objects autonomously
- Powered by Claude Opus 4.x — apply the same anti-over-engineering guard: "Only make changes directly requested..."
- Skills system: markdown-based system prompts loaded via `cortex skill add` — reference skill capabilities rather than re-explaining them
- Snowflake-native: `snowflake_sql_execute` tool for SQL, `st.connection("snowflake")` for Streamlit in Snowflake apps — always prefer these over raw connectors
- Stop conditions and human review triggers are critical — same runaway-loop risk as Claude Code
- For complex tasks: use `cortex ctx task add` / `cortex ctx step add` to break work into tracked steps — the agent loses coherence on long unstructured tasks
- Headless mode (`cortex -p "prompt" --output-format stream-json`) available for CI/automation — output is JSON events, not plain text

---

**Antigravity (Google's agent-first IDE, powered by Gemini 3 Pro)**
- Task-based prompting — describe outcomes, not steps
- Prompt for an Artifact (task list, implementation plan) before execution so you can review it first
Expand Down