惯性聚合 高效追踪和阅读你感兴趣的博客、新闻、科技资讯
阅读原文 在惯性聚合中打开

推荐订阅源

D
DataBreaches.Net
F
Fortinet All Blogs
D
Docker
让小产品的独立变现更简单 - ezindie.com
让小产品的独立变现更简单 - ezindie.com
WordPress大学
WordPress大学
罗磊的独立博客
Y
Y Combinator Blog
Cyber Security Advisories - MS-ISAC
Cyber Security Advisories - MS-ISAC
J
Java Code Geeks
T
The Blog of Author Tim Ferriss
U
Unit 42
N
Netflix TechBlog - Medium
OSCHINA 社区最新新闻
OSCHINA 社区最新新闻
V
V2EX
云风的 BLOG
云风的 BLOG
钛媒体:引领未来商业与生活新知
钛媒体:引领未来商业与生活新知
T
Tailwind CSS Blog
Hugging Face - Blog
Hugging Face - Blog
Stack Overflow Blog
Stack Overflow Blog
爱范儿
爱范儿
酷 壳 – CoolShell
酷 壳 – CoolShell
P
Proofpoint News Feed
G
Google Developers Blog
H
Help Net Security

Hacker News: Show HN

PurrrrrFocus: Pomodoro Timer App - App Store Workflow Engine — Multi-Step Orchestration for Bun RapidPhoto: Pro Photo Editor App - App Store GitHub - DheerG/swarms: Achieve extraordinary results with claude code across a variety of tasks SPICE simulation → oscilloscope → verification with Claude Code — Lucas Gerads Show HN: VCoding – A 5 MB native Windows IDE with no dynamic dependencies Show HN: LLMs don't hallucinate because they're bad at math, it's the format GitHub - Agent-FM/agentfm-core: AgentFM is a peer-to-peer network that turns everyday computers into a decentralized AI supercomputer. AgentFM lets you run massive AI workloads directly across a global mesh of idle CPUs and GPUs. Show HN: Tracking Top US Science Olympiad Alumni over Last 25 Years GitHub - Potarix/agent-hub: One place to talk to all your agents Show HN: Runtime security for AI agents(injection,tool abuse, data exfiltration) GitHub - dubeyKartikay/lazyspotify: Terminal Spotify client for macOS and Linux GitHub - the-banana-tool/king-louie: Easy to use GUI Personal AI Assistant. Win/Linux/Mac. Show HN I made my vacation rental bookable by AI agents–no Airbnb, 0% commission GitHub - basteez/jsf-autoreload: maven plugin to enable hot reload on jsf projects uvm32/hosts/host-gdbstub at main · ringtailsoftware/uvm32 GitHub - labsai/EDDI: Config-driven engine that turns JSON into production-grade AI agents. Multi-agent orchestration, 12+ LLM providers, MCP/A2A protocols, RAG, persistent memory, and enterprise compliance (EU AI Act, GDPR, HIPAA). Built on Quarkus. GitHub - glitchnsec/fortyone-oss: AI Executive Assistant Platform Quickstart | Alien GitHub - muxshed/shed: One stream in, or many. Every destination, simultaneously. No cloud middleman, no per-channel fees, no limits. GitHub - ocrbase-hq/ocrbase: 📄 PDF/IMG ->.MD/JSON Document OCR API for PaddleOCR and GLMOCR. Self-hostable. GitHub - impactjo/home-memory: MCP server that lets your AI assistant remember everything about your home. GitHub - Sets88/dbcls: DbCls is a powerful terminal database client that supports various databases GitHub - neptun2000/heor-agent-mcp GitHub - SeanFDZ/macmind: Single-layer transformer in HyperTalk for the classic Macintosh RollQuation: Math Puzzles - Apps on Google Play GitHub - dropbox/witchcraft Show HN: Agent-cache – Multi-tier LLM/tool/session caching for Valkey and Redis GitHub - opentalon/opentalon: OpenTalon is an open-source platform built from the ground up in Go as a robust alternative to OpenClaw LinkedIn™ 职位抓取工具 - Chrome 应用商店
GitHub - jonverrier/RipStop: Git hook and CI guardrails f...
Jonverrier · 2026-05-12 · via Hacker News: Show HN

Git hook and CI guardrails for AI-assisted software development.

Ripstop is a TypeScript CLI that runs policy checks at Git boundaries (commit, commit message, push, rebase, CI). It targets repos where Cursor, Claude Code, Codex, Amazon Q, or humans make changes: the same rules apply no matter who produced the diff.

It does not replace sandboxing, server-side branch protection, secret scanning, or code review. It does give you a consistent, repo-local enforcement layer, optional agent-readable summaries (RIPSTOP.md), and forensics so casual mistakes and “quietly weaken the guardrails” edits are much harder.

Invoke the CLI as ripstop (for example npx ripstop when the package is a devDependency).


Built-in checks (today)

Each check is configured under checks.<name> in .guardrails.yaml, with mode: off | warn | enforce, triggers, and check-specific options. Built-in presets (for example @jonverrier/ripstop/presets/internal-tooling) wire defaults; repos can extend or override.

Check What it enforces Typical triggers
pii Common PII patterns in files you commit (with exemptions). pre-commit, ci
path-guard Changes under protected globs need an approval trailer in the final commit message (e.g. CHANGE-APPROVED: TICKET-123). commit-msg, ci
test-skip New or disallowed test-skip / disabled-test patterns; optional ticket requirement. pre-commit, ci
history-guard Force-push and remote branch delete on protected branch patterns. pre-push only (Git supplies ref updates on stdin)
ripstop-md-fresh Committed RIPSTOP.md exists and its embedded config hash matches the resolved .guardrails.yaml (including preset merge). pre-commit, ci
reflog-witness Appends witness JSONL per run (branch, .guardrails.yaml hash and optional content on change, RIPSTOP.md hash, etc.). pre-commit, pre-push, pre-rebase, ci

Further behaviour is in docs/ripstop-spec.md. Checks specified but not implemented yet are listed in docs/ripstop-roadmap-plan.md.

Use commit-msg for path-guard (and other trailer-based rules): Git has not finalized the message at pre-commit.


Agent context (0.1.xRIPSTOP.md)

From 0.1.x onward, Ripstop can mirror resolved policy into Layer 1 (agent static context) so agents see active rules before they edit.

  • ripstop generate-md — Writes RIPSTOP.md from merged config (local YAML + extends: presets). Commit it and reference it from AGENTS.md / harness manifests (docs/per-agent-config.md).
  • Per-agent formats--format claude, cursor, codex, q adjust framing; the config hash in the file is the same across formats so ripstop-md-fresh stays consistent.
  • ripstop-md-fresh — Fails when RIPSTOP.md is missing or stale vs resolved config.
  • Flags--output, --check-fresh, --dry-run, --config.

Design notes: docs/ripstop-markdown-enhancement-spec.md.

npx ripstop generate-md
git add RIPSTOP.md

Self-protection (0.2.x)

0.2.x closes the gap where an agent could weaken .guardrails.yaml or generated docs in the same commit as “fixing” a failure.

  1. path-guard preset defaults — Presets add .guardrails.yaml, .guardrails/**, RIPSTOP.md, and .claude/settings*.json (including settings.ripstop.json) to protected_paths, with the same approval trailer as your other paths. Dedicated findings call out guardrails self-protection.
  2. RIPSTOP.md — Generated copy includes what you must not modify; optional Claude output adds harness-oriented deny rules.
  3. ripstop generate-md --format claude — Writes .claude/settings.ripstop.json; merge permissions.deny into your real .claude/settings.json per docs/per-agent-config.md.
  4. reflog-witness — Witness lines include .guardrails.yaml and RIPSTOP.md hashes (and yaml content on change, within limits).
  5. ripstop recover --config-history — Prints witness history (optional --since, --config).

This is not tamper-proof against shell access and --no-verify. Details: docs/ripstop-self-protection-enhancement-spec.md.

npx ripstop recover --config-history
npx ripstop recover --config-history --since 2026-01-01T00:00:00.000Z

Observability

  • Findings — human or JSON (configurable).
  • reporting.audit_log — findings / bypasses (default .git/ripstop/audit.jsonl).
  • reporting.witness_log — witness events (default .git/ripstop/witness.jsonl), including reflog-witness.

Documentation

Document Role
docs/ripstop-spec.md Full product specification.
docs/ripstop-consumer-playbook.md Adoption, hooks, governance, recovery.
docs/per-agent-config.md Per-agent RIPSTOP.md wiring; Claude settings merge.
docs/ripstop-roadmap-plan.md Shipped vs planned.
docs/ripstop-markdown-enhancement-spec.md Design notes for generate-md / ripstop-md-fresh.
docs/ripstop-self-protection-enhancement-spec.md Design notes for self-protection.

Install

npm install --save-dev @jonverrier/ripstop

Example Husky wiring:

npx ripstop check --staged --trigger pre-commit
npx ripstop check --staged --trigger commit-msg --commit-msg-file "$1"
npx ripstop check --trigger pre-push
npx ripstop check --trigger pre-rebase

Minimal .guardrails.yaml

repo:
  name: my-service
  domain: tooling
  tier: 2

extends: "@jonverrier/ripstop/presets/internal-tooling"

checks:
  path-guard:
    mode: enforce
    triggers: [commit-msg, ci]
    protected_paths:
      - "infra/**"
      - "migrations/**"
    approval_trailer: "CHANGE-APPROVED"

  test-skip:
    mode: warn
    triggers: [pre-commit, ci]

The internal-tooling preset already includes self-protection paths, reflog-witness, ripstop-md-fresh, and history-guard. See src/presets/ for the full merged defaults.


CLI

ripstop check [--staged | --all | --diff <ref>] --trigger <trigger> [options]
ripstop generate-md [options]
ripstop recover --config-history [--since <iso>] [--config <path>]
ripstop list
ripstop explain <check> [--resolved]
ripstop version

--trigger: pre-commit, commit-msg, pre-push, pre-rebase, pre-action, ci.


Build

npm install
npm run build
npm run test:ci
npm pack --dry-run

Binary distribution

The package is designed to support Bun-compiled standalone binaries in a later milestone. Binary consumers can use built-in checks and YAML config. npm plugin checks and repo-local TypeScript checks expect a Node-compatible runtime.

License

MIT