← ClaudeAtlas

lint-rule-generatorlisted

Routes, designs, evaluates, and refines deterministic rules from concrete anti-patterns. Use when creating repository policy or a Python, TypeScript, Markdown, SQL, IaC, or config rule; checking whether an upstream rule already covers a problem; calibrating false positives on local corpora; or deciding whether a new rule is safe to introduce as a warning or error.
sarj-ai/code-standards · ★ 5 · Code & Development · score 70
Install: claude install-skill sarj-ai/code-standards
# Lint rule generator Turn a concrete defect description into the narrowest deterministic rule that earns trust on real code. Use `sarj_standards.libs.rules` for problem, catalog, evaluation, and report contracts. Use `sarj_standards.libs.corpus` for manifests, snapshots, local pin verification, and redacted reporting. Keep all executable logic in the uv package; this skill contains no scripts. ## Admit the proposal before routing Before routing or editing any proposed CI guard, preflight, validator, policy workflow, or new or expanded lint rule, record all of the following internally: 1. the concrete failure mode, evidenced by a local incident, a reproducible upstream incident, an authoritative security or correctness advisory, or a minimal reproduction rather than an unsupported hypothetical; 2. the exact existing owner or control for the invariant; 3. the demonstrated gap in that owner or control; 4. the concrete harm if no change is made; 5. whether the proposal distinguishes invalid state from merely changing today's valid configuration; and 6. the smallest fix at the owning boundary. If evidence for the failure mode, demonstrated gap, or no-change harm is absent, or the proposal only freezes today's source spelling, configuration value, or temporary state, stop with `delete/no PR` and ask for an explicit user override. Deleting a redundant check is a valid remediation. Moving the same weak oracle into a script, changing it to inspect plan JSON without prov