verilog-sv-language
a5c-ai/babysitter · Agent Skill
Generates synthesizable RTL code in Verilog and SystemVerilog, adhering to IEEE 1800 standards and best practices.
Jump to install ↓Overview
| Publisher | |
|---|---|
| Repository | a5c-ai/babysitter ↗ |
| Skill folder | verilog-sv-language |
| Stars | 1,747 |
| Bundled files | 1 |
| License | MIT |
| Last updated | May 6, 2026 |
| Links | Skill source ↗ · Repository ↗ |
Mozgi check
What we found in the skill files. Details come with a subscription.
Installation
A skill is a folder with a SKILL.md file. Pick your agent and run the commands in a terminal (macOS, Linux or Git Bash on Windows) from your project root.
Copy the prompt into Claude, Codex or another assistant: it will review the skill and install it for you.
Install the “verilog-sv-language” skill from mozgi.io (https://mozgi.io/skills/verilog-sv-language/) for me. 1. Download the archive: https://mozgi.io/skills/verilog-sv-language/download If it fails, take the skill folder from the source: https://github.com/a5c-ai/babysitter/tree/main/library/specializations/fpga-programming/skills/verilog-sv-language 2. Read SKILL.md and the other files. Check the skill is safe: no hidden commands, no sending data elsewhere, no scripts doing something other than described. If anything looks suspicious, stop and tell me. 3. Put the “verilog-sv-language” folder into your agent's skills folder: .claude/skills/ for Claude Code, .agents/skills/ for Codex, .cursor/skills/ for Cursor, .gemini/skills/ for Gemini CLI. Inside the project if we are in one, otherwise in the home folder (~/.claude/skills/ etc.). 4. Make sure <skills folder>/verilog-sv-language/SKILL.md exists. 5. Briefly tell me what this skill does and offer to try it on my task. If you cannot do a step yourself, tell me what to click or type.
git clone --depth 1 --branch main https://github.com/a5c-ai/babysitter.git /tmp/babysitter mkdir -p .claude/skills cp -r /tmp/babysitter/library/specializations/fpga-programming/skills/verilog-sv-language .claude/skills/verilog-sv-language
Restart Claude Code — the skill appears in the list; call it manually with /verilog-sv-language. For all projects use ~/.claude/skills.
git clone --depth 1 --branch main https://github.com/a5c-ai/babysitter.git /tmp/babysitter mkdir -p .agents/skills cp -r /tmp/babysitter/library/specializations/fpga-programming/skills/verilog-sv-language .agents/skills/verilog-sv-language
Codex looks for skills in .agents/skills from the current folder up to the repo root; call one with /skills or $verilog-sv-language. For all projects use ~/.agents/skills.
git clone --depth 1 --branch main https://github.com/a5c-ai/babysitter.git /tmp/babysitter mkdir -p .cursor/skills cp -r /tmp/babysitter/library/specializations/fpga-programming/skills/verilog-sv-language .cursor/skills/verilog-sv-language
Cursor picks the skill up on start; call it with “/” in Agent chat. Cursor also reads .claude/skills and .agents/skills. For all projects use ~/.cursor/skills.
git clone --depth 1 --branch main https://github.com/a5c-ai/babysitter.git /tmp/babysitter mkdir -p .gemini/skills cp -r /tmp/babysitter/library/specializations/fpga-programming/skills/verilog-sv-language .gemini/skills/verilog-sv-language
Gemini CLI offers to activate the skill when a task matches — you approve it. For all projects use ~/.gemini/skills.
Read SKILL.md and any companion files before deciding whether to install. Click a file to preview it.
SKILL.md contents
Original author text · read-onlyVerilog/SystemVerilog Language Skill
Expert skill for Verilog and SystemVerilog development following IEEE 1364 and IEEE 1800 standards. Provides deep expertise in synthesizable RTL code generation, proper construct usage, and modern coding practices.
Overview
The Verilog/SystemVerilog Language skill enables comprehensive HDL development for FPGA and ASIC designs, supporting:
- IEEE 1800-2017 SystemVerilog standard
- Verilog-2005 backward compatibility
- Proper always_ff, always_comb, always_latch usage
- SystemVerilog interfaces and modports
- Parameterized modules with localparam
- Packed and unpacked arrays
- Packages and imports
Capabilities
1. Proper Always Block Usage
Use SystemVerilog always block variants correctly:
// Sequential logic - always_ff
always_ff @(posedge clk or negedge rst_n) begin
if (!rst_n) begin
counter <= '0;
state <= IDLE;
end else begin
counter <= counter + 1'b1;
state <= next_state;
end
end
// Combinational logic - always_comb
always_comb begin
// Default assignments prevent latches
next_state = state;
output_valid = 1'b0;
case (state)
IDLE: begin
if (start) next_state = RUN;
end
RUN: begin
output_valid = 1'b1;
if (done) next_state = IDLE;
end
default: next_state = IDLE;
endcase
end
// Intentional latch - always_latch (rare)
always_latch begin
if (enable)
latch_out = data_in;
end
2. Parameterized Modules
Create reusable parameterized modules:
module sync_fifo #(
parameter int DATA_WIDTH = 8,
parameter int DEPTH = 16,
parameter int ALMOST_FULL_THRESH = DEPTH - 2,
parameter int ALMOST_EMPTY_THRESH = 2,
// Derived parameters using localparam
localparam int ADDR_WIDTH = $clog2(DEPTH),
localparam int CNT_WIDTH = $clog2(DEPTH + 1)
) (
input logic clk,
input logic rst_n,
// Write interface
input logic wr_en,
input logic [DATA_WIDTH-1:0] wr_data,
output logic full,
output logic almost_full,
// Read interface
input logic rd_en,
output logic [DATA_WIDTH-1:0] rd_data,
output logic empty,
output logic almost_empty,
// Status
output logic [CNT_WIDTH-1:0] fill_level
);
// Memory array
logic [DATA_WIDTH-1:0] mem [DEPTH];
// Pointers
logic [ADDR_WIDTH-1:0] wr_ptr, rd_ptr;
logic [CNT_WIDTH-1:0] count;
// Write logic
always_ff @(posedge clk or negedge rst_n) begin
if (!rst_n) begin
wr_ptr <= '0;
end else if (wr_en && !full) begin
mem[wr_ptr] <= wr_data;
wr_ptr <= wr_ptr + 1'b1;
end
end
// Read logic
always_ff @(posedge clk or negedge rst_n) begin
if (!rst_n) begin
rd_ptr <= '0;
end else if (rd_en && !empty) begin
rd_ptr <= rd_ptr + 1'b1;
end
end
// Read data (registered output)
always_ff @(posedge clk) begin
if (rd_en && !empty) begin
rd_data <= mem[rd_ptr];
end
end
// Count logic
always_ff @(posedge clk or negedge rst_n) begin
if (!rst_n) begin
count <= '0;
end else begin
case ({wr_en && !full, rd_en && !empty})
2'b10: count <= count + 1'b1;
2'b01: count <= count - 1'b1;
default: count <= count;
endcase
end
end
// Status outputs
assign full = (count == DEPTH);
assign empty = (count == '0);
assign almost_full = (count >= ALMOST_FULL_THRESH);
assign almost_empty = (count <= ALMOST_EMPTY_THRESH);
assign fill_level = count;
endmodule
3. SystemVerilog Interfaces
Define reusable interfaces with modports:
// AXI-Stream interface definition
interface axis_if #(
parameter int DATA_WIDTH = 32,
parameter int USER_WIDTH = 1,
parameter int ID_WIDTH = 1
) (
input logic aclk,
input logic aresetn
);
logic tvalid;
logic tready;
logic [DATA_WIDTH-1:0] tdata;
logic [DATA_WIDTH/8-1:0] tstrb;
logic [DATA_WIDTH/8-1:0] tkeep;
logic tlast;
logic [ID_WIDTH-1:0] tid;
logic [ID_WIDTH-1:0] tdest;
logic [USER_WIDTH-1:0] tuser;
// Master modport
modport master (
input aclk, aresetn, tready,
output tvalid, tdata, tstrb, tkeep, tlast, tid, tdest, tuser
);
// Slave modport
modport slave (
input aclk, aresetn, tvalid, tdata, tstrb, tkeep, tlast, tid, tdest, tuser,
output tready
);
// Monitor modport for verification
modport monitor (
input aclk, aresetn, tvalid, tready, tdata, tstrb, tkeep, tlast, tid, tdest, tuser
);
// Helper tasks for verification
task automatic wait_for_handshake();
@(posedge aclk);
while (!(tvalid && tready)) @(posedge aclk);
endtask
endinterface
// Using the interface in a module
module axis_register #(
parameter int DATA_WIDTH = 32
) (
input logic clk,
input logic rst_n,
axis_if.slave s_axis,
axis_if.master m_axis
);
// Skid buffer implementation
logic [DATA_WIDTH-1:0] data_reg;
logic valid_reg;
always_ff @(posedge clk or negedge rst_n) begin
if (!rst_n) begin
valid_reg <= 1'b0;
data_reg <= '0;
end else if (s_axis.tready) begin
valid_reg <= s_axis.tvalid;
data_reg <= s_axis.tdata;
end
end
assign m_axis.tvalid = valid_reg;
assign m_axis.tdata = data_reg;
assign s_axis.tready = m_axis.tready || !valid_reg;
endmodule
4. Packages and Imports
Create and use SystemVerilog packages:
// Package definition
package fpga_pkg;
// Type definitions
typedef enum logic [2:0] {
IDLE = 3'b000,
INIT = 3'b001,
RUN = 3'b010,
PAUSE = 3'b011,
DONE = 3'b100,
ERROR = 3'b101
} state_t;
// Struct definitions
typedef struct packed {
logic valid;
logic [31:0] data;
logic [3:0] strb;
logic last;
} axi_data_t;
// Constants
localparam int CLK_FREQ_HZ = 100_000_000;
localparam int TIMEOUT_CYCLES = CLK_FREQ_HZ / 1000; // 1ms
// Functions
function automatic int clog2(int value);
int result = 0;
value = value - 1;
while (value > 0) begin
result++;
value = value >> 1;
end
return result;
endfunction
function automatic logic [31:0] reverse_bits(logic [31:0] data);
for (int i = 0; i < 32; i++) begin
reverse_bits[i] = data[31-i];
end
endfunction
endpackage
// Using the package
module my_design
import fpga_pkg::*;
(
input logic clk,
input logic rst_n,
input logic start,
output state_t current_state,
output axi_data_t output_data
);
state_t state, next_state;
always_ff @(posedge clk or negedge rst_n) begin
if (!rst_n)
state <= IDLE;
else
state <= next_state;
end
always_comb begin
next_state = state;
case (state)
IDLE: if (start) next_state = RUN;
RUN: next_state = DONE;
DONE: next_state = IDLE;
default: next_state = IDLE;
endcase
end
assign current_state = state;
endmodule
5. Packed and Unpacked Arrays
Use arrays correctly for synthesis:
module array_examples #(
parameter int WIDTH = 8,
parameter int DEPTH = 4
) (
input logic clk,
input logic rst_n,
input logic [WIDTH-1:0] data_in,
output logic [WIDTH-1:0] data_out
);
// Unpacked array - multiple memory locations
logic [WIDTH-1:0] memory_array [DEPTH]; // Memory inference
// Packed array - single contiguous bit vector
logic [DEPTH-1:0][WIDTH-1:0] shift_reg; // Shift register
// Multi-dimensional packed array
logic [3:0][7:0] packed_data; // 32-bit value as 4 bytes
// Multi-dimensional unpacked array
logic [7:0] mem_2d [4][8]; // 4x8 array of bytes
// Shift register using packed array
always_ff @(posedge clk or negedge rst_n) begin
if (!rst_n) begin
shift_reg <= '0;
end else begin
shift_reg <= {shift_reg[DEPTH-2:0], data_in};
end
end
// Memory write
logic [$clog2(DEPTH)-1:0] wr_addr;
always_ff @(posedge clk) begin
memory_array[wr_addr] <= data_in;
end
// Byte access to packed data
assign packed_data[0] = 8'hAB; // Least significant byte
assign packed_data[3] = 8'hCD; // Most significant byte
assign data_out = shift_reg[DEPTH-1];
endmodule
6. Blocking vs Non-Blocking Assignments
Apply assignments correctly:
// CORRECT: Non-blocking for sequential logic
always_ff @(posedge clk) begin
reg_a <= data_in; // Non-blocking
reg_b <= reg_a; // Non-blocking - creates pipeline
reg_c <= reg_b; // Non-blocking
end
// CORRECT: Blocking for combinational logic
always_comb begin
temp = a & b; // Blocking
result = temp | c; // Blocking - uses updated temp
end
// INCORRECT: Don't mix in sequential blocks
always_ff @(posedge clk) begin
temp = data_in; // WRONG: blocking in sequential
reg_a <= temp;
end
// CORRECT: Use intermediate signals
logic temp_comb;
always_comb begin
temp_comb = data_in & enable;
end
always_ff @(posedge clk) begin
reg_a <= temp_comb; // Non-blocking
end
7. Synthesis Attributes
Apply vendor-specific attributes:
module synthesis_attributes (
input logic clk,
input logic rst_n,
input logic async_in,
output logic sync_out
);
// Xilinx: ASYNC_REG for synchronizers
(* ASYNC_REG = "TRUE" *) logic [1:0] sync_reg;
// Xilinx: Keep signal for debugging
(* KEEP = "TRUE" *) logic debug_signal;
// Xilinx: RAM style control
(* RAM_STYLE = "block" *) logic [7:0] block_mem [1024];
(* RAM_STYLE = "distributed" *) logic [7:0] dist_mem [16];
// Xilinx: FSM encoding
(* FSM_ENCODING = "one_hot" *) enum logic [2:0] {
IDLE, RUN, DONE
} state;
// Intel: RAM inference
(* ramstyle = "M20K" *) logic [7:0] intel_mem [1024];
// Synchronizer implementation
always_ff @(posedge clk or negedge rst_n) begin
if (!rst_n) begin
sync_reg <= '0;
end else begin
sync_reg <= {sync_reg[0], async_in};
end
end
assign sync_out = sync_reg[1];
endmodule
Process Integration
This skill integrates with the following processes:
| Process | Integration Point |
|---------|-------------------|
| verilog-systemverilog-design.js | Primary Verilog/SV development |
| testbench-development.js | SV testbench generation |
| rtl-module-architecture.js | Module design |
| uvm-testbench.js | UVM environment development |
Output Schema
{
"module": {
"name": "sync_fifo",
"language": "SystemVerilog",
"standard": "IEEE 1800-2017",
"parameters": [
{ "name": "DATA_WIDTH", "type": "int", "default": 8 },
{ "name": "DEPTH", "type": "int", "default": 16 }
],
"ports": [
{ "name": "clk", "direction": "input", "type": "logic" },
{ "name": "rst_n", "direction": "input", "type": "logic" }
],
"interfaces": []
},
"package": {
"name": "fpga_pkg",
"types": ["state_t", "axi_data_t"],
"functions": ["clog2", "reverse_bits"],
"constants": ["CLK_FREQ_HZ", "TIMEOUT_CYCLES"]
},
"compliance": {
"synthesizable": true,
"alwaysBlocksCorrect": true,
"noInferredLatches": true,
"noBlockingInSequential": true
},
"artifacts": [
"src/sync_fifo.sv",
"src/fpga_pkg.sv",
"src/axis_if.sv",
"tb/sync_fifo_tb.sv"
]
}
Best Practices
Always Block Selection
- Use
always_fffor flip-flops (sequential logic) - Use
always_combfor combinational logic - Use
always_latchonly for intentional latches (rare) - Never use plain
always @*in SystemVerilog
Assignment Rules
- Non-blocking (
<=) inalways_ff - Blocking (
=) inalways_comb - Never mix assignment types in same block
Parameterization
- Use
parameterfor configurable values - Use
localparamfor derived/internal constants - Use
$clog2()for address width calculation - Provide sensible defaults
Style Guidelines
- One port per line for readability
- Use
logicinstead ofwire/reg - Use
'0and'1for all-zeros/all-ones - Use meaningful names (snake_case preferred)
References
- IEEE Std 1800-2017 (SystemVerilog)
- IEEE Std 1364-2005 (Verilog)
- Verilator User Manual
- Xilinx UG901: Vivado Synthesis Guide
- Intel Quartus Prime Synthesis Guide
See Also
verilog-systemverilog-design.js- Verilog/SV development processuvm-testbench.js- UVM testbench development- SK-001: VHDL Language skill
- SK-003: SystemVerilog Assertions (SVA) skill
- AG-001: RTL Design Expert agent
Frequently asked questions
What does the verilog-sv-language skill do?
Generates synthesizable RTL code in Verilog and SystemVerilog, adhering to IEEE 1800 standards and best practices.
How do I install verilog-sv-language?
Open the “Installation” block on this page and pick your agent — Claude Code, Codex, Cursor or Gemini CLI — for ready-made commands that copy the skill folder to the right place. Or copy the prompt from the “Via assistant” tab and let the assistant review and install the skill.
Which AI models work with verilog-sv-language?
Skills are plain Markdown instructions (the Agent Skills format), so verilog-sv-language works with any agent that reads SKILL.md: Claude Code, Codex, Cursor, Gemini CLI, GitHub Copilot, OpenCode and more. The model loads the instructions only when a task matches the skill's description.
Is verilog-sv-language free?
The skill is published openly by a5c-ai under the MIT license. The skill itself costs nothing — you only pay your AI provider for the tokens you use.
Is it safe to install skills?
A skill is a set of instructions your agent will follow, so read SKILL.md and its companion files before installing. Scripts in the skill folder only run if your agent runs them — keep an eye on what it does and do not grant unnecessary permissions.
How is a skill different from an MCP server or a plugin?
An MCP server or plugin gives the model new tools — code that runs and returns a result. A skill gives it knowledge and process: how to approach a task, which steps to follow and what good output looks like. It needs no server or API key.