Stats
Actions
Tags
From fpga-lsp
Run Verible's lint over SystemVerilog/Verilog files and surface diagnostics. Use when the user wants lint output for one or more `.sv`/`.svh`/`.v`/`.vh` files (e.g. "lint this file", "what does Verible say about X", "check this RTL for warnings").
How this skill is triggered — by the user, by Claude, or both
Slash command
/fpga-lsp:sv-lintThis skill is limited to the following tools:
The summary Claude sees in its skill listing — used to decide when to auto-load this skill
Thin wrapper over `verible-verilog-lint`. Invoke the CLI, return the diagnostics, let the caller interpret them.
Thin wrapper over verible-verilog-lint. Invoke the CLI, return the diagnostics, let the caller interpret them.
Single file:
verible-verilog-lint --rules_config_search <path>
Many files: pass each path as a positional argument in one invocation:
verible-verilog-lint --rules_config_search <path1> <path2> ...
Always pass --rules_config_search so a project-level .rules.verible_lint (walking up from each source file) is respected without forcing the user to point at it.
If a workspace-root verible.filelist exists, or a plugin-managed filelist exists at ${CLAUDE_PLUGIN_DATA}/filelists/<workspace-hash>.filelist, optionally pass --file_list_path=<that path> for cross-file rules. Prefer the project file when both exist.
Verible emits one diagnostic per line:
<path>:<line>:<col>: <message> [<rule-id>]
<rule-id> is the canonical Verible rule (e.g. always-comb, case-missing-default, explicit-parameter-storage-type). Cite it verbatim when reporting.[rule-id] are usually parser errors — surface them as parse errors, not lint warnings.When multiple paths are passed, group diagnostics by file in the response. Keep rule IDs intact so the caller can look them up. Do not invent severities Verible did not emit.
This skill SHALL invoke only verible-verilog-lint. No bespoke checks, no rewriting of messages, no interpretation beyond grouping.
npx claudepluginhub nyavana/fpga-lsp --plugin fpga-lspCreates, edits, and optimizes skills for Claude Code, including drafting, evaluating with test prompts, iterating on performance, and improving skill descriptions for better triggering accuracy.