Claude Linter — Development skill for Claude Code
Lints Claude Code config, CLAUDE.md, hooks, MCP, settings, permissions for silent conflicts, dead references, and token bloat.
How to install Claude Linter
This entry records only its repository, not the path inside it, so there is no
exact command to give. Open Santhosraj/Claude_Linter and copy the folder into
~/.claude/skills/, or the file into ~/.claude/agents/.
What Claude Linter does
Lints Claude Code config, CLAUDE.md, hooks, MCP, settings, permissions for silent conflicts, dead references, and token bloat. Offline by default.
Alternatives in Development
- Claude Config — Built-in /config command to manage settings interactively 81.1k ★
- Resolve Drizzle Migrations — Resolve Drizzle migration conflicts by accepting upstream and regenerating local migrations 2.5k ★
- Forge Stats — Track command sizes, detect bloat, compare against baselines and thresholds 117 ★
README
cclint
[](https://www.npmjs.com/package/@santhosraj/cclint) [](https://github.com/Santhosraj/Claude_Linter/actions/workflows/ci.yml) [](https://github.com/Santhosraj/Claude_Linter/actions/workflows/action-selftest.yml) [](LICENSE)
Lint the configuration a Claude Code project accumulates — `CLAUDE.md`, hooks, MCP servers — for the failures that are silent until the agent starts behaving oddly: dead references, config that a higher layer quietly overrides, instructions that contradict each other, and context you're paying for on every turn without realising it.
npx @santhosraj/cclint@latest --strict
Real output, on a project with three ordinary mistakes in it:
.claude\settings.json
4:59 warn Hook command path does not exist: .claude/hooks/guard.sh hooks/dead-command
│ Checked relative to the project root.
8:15 warn 2 `permissions.allow` entries are ignored: this workspace has not been trusted. permissions/untrusted-workspace
│ Claude Code drops project-level allow entries until the workspace is trusted.
│ Run Claude Code interactively here once and accept the trust dialog, or set
│ projects["D:/tmp/api"].hasTrustDialogAccepted: true
│ in D:/tmp/api/.home/.claude.json
│ `deny` and `ask` are unaffected, as is your user-level allow list.
│ Both project files are gated: settings.json and settings.local.json.
CLAUDE.md
8:1 info Possible conflict on indentation style: "tabs" vs "spaces". memory/axis-conflict
│ CLAUDE.md:4 — Use tabs for indentation in this repo.
│ CLAUDE.md:8 — Ne
Related Skills
Hooklint
Offline linter for the agent config that actually executes: hooks, skills, MCP entries, permission rules. Catc
Claude Config Transfer
Two PowerShell scripts that move a full Claude Code setup (settings, permissions, personal skills, all marketp
Dead Code
List orphan functions, classes, atoms, and endpoints with no inbound references — framework entry points exclu
Massu Cleanup
Dead code removal — unused imports, orphaned files, dead exports, and stale references
BDD Drift
Health-check живой документации BDD: dead references, coverage gaps, trivially-green сценарии. Usage — /bdd-dr
Claude Config Auditor
Non-destructive linter and cost auditor for Claude Code's .claude/ and CLAUDE.md — measures token cost, audits
Related Agents
Claude Code Infra
Use PROACTIVELY and automatically — do not wait to be asked — to configure Claude Code itself: sub-agents, ski
Skill Linter
Lints the SKILL.md files in a skills folder for broken frontmatter, a name that does not match its folder, mis
Loom Hermit
Loom Hermit - Simplification specialist that identifies and proposes removal of unnecessary complexity, bloat,