Releases·Tool Release
Dev launches AI Guard, an ESLint plugin catching AI-assistant code patterns — Reddit feedback reshapes the roadmap
An independent developer launched AI Guard on Product Hunt, an MIT-licensed open-source ESLint plugin with 18 deterministic rules targeting error patterns AI coding assistants introduce in JS/TS. It ships as a CLI, a GitHub Action with SARIF output, and an init-context command that generates CLAUDE.md/.cursorrules files. Launch feedback is already reshaping the roadmap toward measuring whether instruction files actually change agent behavior.
3 engagement1 sourceFri, Sep 18, 2026, 02:58 PM
AI Overview
- AI Guard ships 18 deterministic ESLint rules catching AI-assistant patterns in JS/TS: floating promises, empty catch blocks, hardcoded secrets, SQL string concat, missing authz checks
- Distribution includes a CLI, a GitHub Action with SARIF output, and init-context generating CLAUDE.md / .cursorrules so agents learn rules before writing code
- Creator says no controlled measurement exists yet on whether instruction files change agent behavior, making it the top roadmap item
- Dead-code and duplicate-logic rules stay out of the recommended preset permanently after pushback that humans write such patterns on purpose
- MIT-licensed project prioritizes rule requests and false-positive reports over stars; most engaged technical feedback focused on precision, not coverage
Related
Social·Community
AI Guard maintainer: commenter found flaw in init-context measurement; linter timestamps can't map violations to agent turns
The maintainer of AI Guard, an open-source ESLint plugin for AI-assisted JS/TS, posted a follow-up acknowledging a flaw in its planned benchmark. A commenter, Deep_Ad1959, pointed out the linter fires at CI time stamped with the commit, not the agent turn that wrote the code, so it can't directly measure whether init-context reduces agent violations. The team is redesigning the measurement around post-tool-use hooks.
The maintainer of AI Guard, an open-source ESLint plugin for AI-assisted JS/TS, posted a follow-up acknowledging a flaw in its planned benchmark. A commenter, Deep_Ad1959, pointed out the linter fires at CI time stamped with the commit, not the agent turn that wrote the code, so it can't directly measure whether init-context reduces agent violations. The team is redesigning the measurement around post-tool-use hooks.
7 engagement1 sourceredditFri, Sep 18, 2026, 02:57 PM
Products·launched
AI Guard — 18 ESLint rules catching patterns AI coding assistants introduce in JS/TS
AI Guard is a set of 18 ESLint rules for JavaScript/TypeScript that detect recurring code-quality mistakes AI coding assistants make, such as floating promises, empty catch blocks, hardcoded secrets, and SQL string concatenation. It targets developers who use tools like Claude Code, Cursor, and Copilot and want to stop these patterns from reaching commits. LLMs are the source of the problematic patterns being caught, not part of the tool itself.
AI Guard is a set of 18 ESLint rules for JavaScript/TypeScript that detect recurring code-quality mistakes AI coding assistants make, such as floating promises, empty catch blocks, hardcoded secrets, and SQL string concatenation. It targets developers who use tools like Claude Code, Cursor, and Copilot and want to stop these patterns from reaching commits. LLMs are the source of the problematic patterns being caught, not part of the tool itself.
2 engagement2 sourcesredditMon, Sep 7, 2026, 06:25 AM
Releases·Tool Release
Show HN: Deslop – Oxlint rules that flag unsafe types, duplicate code, and low-value tests in AI-generated TypeScript
A developer launched Deslop on GitHub via Show HN, a set of Oxlint rules targeting TypeScript code written by AI. It flags unsafe types, duplicate code, and low-value tests.
A developer launched Deslop on GitHub via Show HN, a set of Oxlint rules targeting TypeScript code written by AI. It flags unsafe types, duplicate code, and low-value tests.
39 engagement1 sourcehackernewsTue, Oct 13, 2026, 05:52 PM
Social·Community
Developer proposes tool to measure CLAUDE.md file effectiveness for coding agents
A developer suggests building a tool that runs coding agents on tasks twice—once with CLAUDE.md and once without—to compare token usage, test pass rates, and edit counts. The idea aims to quantify whether the configuration file actually improves agent performance.
A developer suggests building a tool that runs coding agents on tasks twice—once with CLAUDE.md and once without—to compare token usage, test pass rates, and edit counts. The idea aims to quantify whether the configuration file actually improves agent performance.
12 engagement2 sourcesredditSat, Jul 18, 2026, 11:42 PM
Products·launched
AI Guard: ESLint plugin and CLI to catch AI code issues
AI Guard is an ESLint plugin and CLI that acts as a safety net for developers using AI coding assistants like Claude Code, Cursor, and Copilot. It automatically detects common problematic patterns in generated code, such as floating promises, empty catch blocks, placeholder secrets, and missing auth middleware, helping developers avoid committing flawed code.
AI Guard is an ESLint plugin and CLI that acts as a safety net for developers using AI coding assistants like Claude Code, Cursor, and Copilot. It automatically detects common problematic patterns in generated code, such as floating promises, empty catch blocks, placeholder secrets, and missing auth middleware, helping developers avoid committing flawed code.
1 engagement1 sourceredditThu, Aug 6, 2026, 02:37 PM