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

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
V
V2EX
WordPress大学
WordPress大学
U
Unit 42
I
InfoQ
A
About on SuperTechFans
宝玉的分享
宝玉的分享
J
Java Code Geeks
博客园 - 司徒正美
爱范儿
爱范儿
Engineering at Meta
Engineering at Meta
G
Google Developers Blog
人人都是产品经理
人人都是产品经理
小众软件
小众软件
Microsoft Security Blog
Microsoft Security Blog
L
LangChain Blog
奇客Solidot–传递最新科技情报
奇客Solidot–传递最新科技情报
Hugging Face - Blog
Hugging Face - Blog
H
Hackread – Cybersecurity News, Data Breaches, AI and More
aimingoo的专栏
aimingoo的专栏
OSCHINA 社区最新新闻
OSCHINA 社区最新新闻
Last Week in AI
Last Week in AI
腾讯CDC
Recent Announcements
Recent Announcements

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 - skaldlab/muninn: 🐦‍⬛ Security scanner for GitHub...
sg0nzalez83 · 2026-06-20 · via Hacker News: Show HN

CI CodeQL Release License: AGPL v3 Go Version

Muninn is an all-in-one open-source security scanner for GitHub Actions pipelines and self-hosted CI, built by Skald Lab.

Named after Odin's raven of Memory from Norse mythology, Muninn remembers every vulnerability it has ever seen — and reports them all back in one unified format.

Muninn orchestrates eight best-in-class open-source scanners, normalizes their output into a single finding schema, and delivers results as GitHub PR comments, SARIF uploads, and structured JSON — with a single uses: line.


Quick Start

Add this to any GitHub Actions workflow:

- name: Muninn Security Scan
  uses: skaldlab/muninn@v0.3.3
  with:
    token: ${{ secrets.GITHUB_TOKEN }}
    fail-on: high

Muninn runs all eight scanners, posts a summary comment on your pull request, uploads SARIF to the GitHub Security tab, and fails the check when findings meet or exceed your fail-on threshold.


What Muninn Scans

Scanner What it finds Upstream
gitleaks Secrets, API keys, and credentials in source gitleaks/gitleaks
zizmor CI/CD pipeline misconfigurations and injection risks woodruffw/zizmor
actionlint GitHub Actions syntax errors and security anti-patterns rhysd/actionlint
poutine Supply chain risks in CI/CD (unpinned actions, risky patterns) boostsecurityio/poutine
semgrep Application SAST (injection, insecure patterns, secrets in code) semgrep.dev
osv-scanner Known CVEs in dependencies (Go, npm, pip, and more) google/osv-scanner
trivy Container image and filesystem vulnerabilities aquasecurity/trivy
checkov Infrastructure-as-Code misconfigurations (Terraform, K8s, Docker) bridgecrewio/checkov

Cross-scanner deduplication

Running multiple dependency scanners means the same CVE can surface more than once — OSV-Scanner flags it in a lockfile while Trivy flags it in a container layer. Muninn collapses these into a single finding keyed on the advisory id (a CVE-… is preferred over GHSA-… so the same vulnerability converges across scanners) scoped to the affected package, so distinct packages sharing a CVE stay separate.

Because an aggregated finding no longer belongs to one tool, dependency findings render under a neutral [dependency] heading (rather than [osv-scanner]) with structured detail, and the scanners are surfaced via attribution instead:

  • PR commentPackage, Advisory (with the shared CVE in parentheses), Detected by, and a Sources list showing where each scanner saw it (e.g. package-lock.json (osv-scanner), node:18 (trivy)).
  • JSON reportdetected_by (the scanner list) and sources (per-scanner tool + file pairs).
  • SARIF — a detectedBy property on the result.

The severity summary counts these aggregated unique findings, not raw scanner hits, because deduplication runs before any report is written.


Configuration

Create a muninn.yml at the root of your repository to customize scanner behavior, severity thresholds, and suppressions.

version: 1

# Minimum severity to fail the run.
# Options: critical | high | medium | low | info
fail-on: critical

scanners:
  gitleaks:
    enabled: true

  zizmor:
    enabled: true

  actionlint:
    enabled: true

  poutine:
    enabled: true

  semgrep:
    enabled: true
    rulesets:
      - p/security-audit
      - p/secrets
    exclude-paths:
      - tests/
      - fixtures/

  osv-scanner:
    enabled: true

  trivy:
    enabled: true
    # Optional: narrow what Trivy reports (default: all levels).
    # severity: [CRITICAL, HIGH]
    ignore-unfixed: true

  checkov:
    enabled: true
    skip-checks: []

# Suppress specific findings. Every entry must include a reason.
suppressions:
  - id: fixtures/
    reason: "Intentional test fixtures, not real secrets or vulnerabilities"
    expires: ""

  - fingerprint: abc123def456
    reason: "Known false positive in generated code"
    expires: "2026-12-31T23:59:59Z"

  - tool: gitleaks
    rule-id: generic-api-key
    reason: "Test fixture file, not a real secret"
    expires: ""

Top-level fields

Field Type Default Description
version int 1 Config schema version. Must be 1.
fail-on string critical Minimum severity that causes a non-zero exit code

Scanner fields

Each key under scanners matches a scanner name. All scanners support enabled (default true).

Scanner Additional fields Description
semgrep rulesets, exclude-paths Semgrep rule packs and path prefixes to skip
trivy severity, ignore-unfixed Trivy severity filter (default: UNKNOWN through CRITICAL; omit to scan all levels) and whether to omit unfixed CVEs
checkov skip-checks Checkov check IDs to skip

Suppression fields

Field Description
id Suppress findings whose file path contains this substring
fingerprint Suppress a specific finding by its Muninn fingerprint
tool Scanner name; with rule-id, both must match. Alone, suppresses all findings from that scanner
rule-id Scanner-native rule identifier; with tool, both must match. Alone, suppresses that rule from any scanner
reason Required. Human-readable justification
expires Optional RFC 3339 UTC timestamp; omit for permanent suppressions

Action inputs

Input Default Description
token ${{ github.token }} GitHub token for PR comments and SARIF upload
fail-on critical Minimum severity to exit non-zero
config muninn.yml Path to config file relative to repository root
format sarif,comment Comma-separated output formats: sarif, json, comment
output muninn.sarif SARIF output path override
target . Path to scan (repository root)

Action outputs

Output Description
findings-count Total non-suppressed findings across all scanners
critical-count Critical-severity finding count
high-count High-severity finding count
medium-count Medium-severity finding count
low-count Low-severity finding count
sarif-path Path to the generated SARIF report
json-path Path to the generated JSON report

Why Muninn vs Snyk, GHAS, or Manual Tools?

Muninn Snyk GitHub Advanced Security Manual tool chain
License AGPL-3.0 (open source) Proprietary SaaS Proprietary (Enterprise or public repos) Mixed
Data leaves your repo No — runs on your runner Yes — code uploaded to Snyk Yes — processed by GitHub Depends on tool
CI/CD pipeline security Yes (zizmor, actionlint, poutine) Limited No Requires assembly
Supply chain / pipeline risks Yes (poutine) Partial No Requires assembly
Self-hostable Yes — Docker image or binary No No Yes, with effort
Unified findings Yes — one schema, one comment Partial Partial No — N different formats
Setup One uses: line Account + integration Org licensing Install and wire each tool

Muninn is not a replacement for every security program — it is a practical default for teams that want broad coverage without stitching together eight separate tools.


CLI and Self-Hosted Usage

Docker (recommended)

The official image bundles Muninn and all eight scanner binaries:

docker run --rm \
  -v "$(pwd):/github/workspace" \
  -w /github/workspace \
  ghcr.io/skaldlab/muninn:0.3.3 \
  --target . \
  --output json,sarif \
  --fail-on high

Binary

Download a release binary from GitHub Releases or install with Go:

go install github.com/skaldlab/muninn@v0.3.3

Scanner binaries (gitleaks, semgrep, checkov, and the rest) must be on PATH. The Docker image includes everything pre-installed.

Verifying releases

Every release is signed with cosign using keyless (OIDC) signing — there are no long-lived keys to trust. The container image also ships with an SBOM and a max-mode SLSA provenance attestation.

Verify the container image:

cosign verify \
  --certificate-identity-regexp '^https://github.com/skaldlab/muninn/\.github/workflows/release\.yml@' \
  --certificate-oidc-issuer 'https://token.actions.githubusercontent.com' \
  ghcr.io/skaldlab/muninn:0.3.3

Verify release binaries via the signed checksums file (download checksums.txt and the Sigstore bundle checksums.txt.sigstore.json from the release):

cosign verify-blob \
  --bundle checksums.txt.sigstore.json \
  --certificate-identity-regexp '^https://github.com/skaldlab/muninn/\.github/workflows/release\.yml@' \
  --certificate-oidc-issuer 'https://token.actions.githubusercontent.com' \
  checksums.txt

# Then confirm the binary matches the verified checksums:
shasum -a 256 -c checksums.txt

Inspect the image's SBOM and SLSA provenance attestations (attached by BuildKit):

docker buildx imagetools inspect ghcr.io/skaldlab/muninn:0.3.3 \
  --format '{{ json .SBOM }}'
docker buildx imagetools inspect ghcr.io/skaldlab/muninn:0.3.3 \
  --format '{{ json .Provenance }}'

CLI flags

Flag Env var Default Description
--config CONFIG_PATH muninn.yml Path to configuration file
--target SCAN_TARGET . Repository root to scan
--fail-on FAIL_ON from config Minimum severity to exit non-zero
--output OUTPUT_FORMATS json Comma-separated formats: json, sarif, comment
--version Print version and exit

Contributing

Contributions are welcome. See CONTRIBUTING.md for scanner development guidelines, test commands, and pull request conventions.


License

Muninn is licensed under the GNU Affero General Public License v3.0. You are free to use, modify, and distribute Muninn — including running it as a service — but modifications must remain open source under AGPL-3.0.

For commercial licensing or enterprise support, contact hello@skaldlab.dev.


Built with 🐦‍⬛ by Skald Lab