Compare commits

...

4 Commits

Author SHA1 Message Date
Gauvain
a51c4868c7 Merge branch 'develop' into ci/trivy-scan 2026-06-05 14:33:01 +02:00
Gauvain
3dbe5bb64c ci(issues): detect likely-duplicate issues on open (#1645)
Some checks are pending
🏗️ Build Apps / 🤖 Build Android APK (Phone) (push) Waiting to run
🏗️ Build Apps / 🤖 Build Android APK (TV) (push) Waiting to run
🏗️ Build Apps / 🍎 Build iOS IPA (Phone) (push) Waiting to run
🏗️ Build Apps / 🍎 Build iOS IPA (Phone - Unsigned) (push) Waiting to run
🏗️ Build Apps / 🍎 Build tvOS IPA (push) Waiting to run
🏗️ Build Apps / 🍎 Build tvOS IPA (Unsigned) (push) Waiting to run
🔒 Lockfile Consistency Check / 🔍 Check bun.lock and package.json consistency (push) Waiting to run
🛡️ CodeQL Analysis / 🔎 Analyze with CodeQL (actions) (push) Waiting to run
🛡️ CodeQL Analysis / 🔎 Analyze with CodeQL (javascript-typescript) (push) Waiting to run
🏷️🔀Merge Conflict Labeler / 🏷️ Labeling Merge Conflicts (push) Waiting to run
🚦 Security & Quality Gate / 🚑 Expo Doctor Check (push) Waiting to run
🚦 Security & Quality Gate / 📝 Validate PR Title (push) Waiting to run
🚦 Security & Quality Gate / 🔍 Vulnerable Dependencies (push) Waiting to run
🚦 Security & Quality Gate / 🔍 Lint & Test (typecheck) (push) Waiting to run
🚦 Security & Quality Gate / 🔍 Lint & Test (check) (push) Waiting to run
🚦 Security & Quality Gate / 🔍 Lint & Test (format) (push) Waiting to run
🚦 Security & Quality Gate / 🔍 Lint & Test (lint) (push) Waiting to run
2026-06-05 14:21:12 +02:00
Gauvino
f5dcc73344 ci(security): scan every push and stabilise the Trivy DB cache key
Drop the push paths filter so secret and misconfig scans cover all file types
(YAML, JSON, native, scripts), not just JS/TS. Replace the per-run
github.run_id cache key with a weekly per-OS key, so the vulnerability DB is
reused within the week instead of writing a fresh immutable cache entry on
every run.
2026-06-05 13:16:13 +02:00
Gauvino
44492876b3 ci(security): add Trivy filesystem scan to code scanning
Streamyfin ships no container image, so this runs a Trivy *filesystem* scan
(vulnerable deps, secrets, misconfig) and uploads SARIF to GitHub code scanning,
complementing CodeQL and dependency-review. Runs on push to develop/master,
weekly, and on demand (not on PRs — fork PRs can't upload SARIF, and
dependency-review already gates PR dependencies). Reports findings (CRITICAL/HIGH,
ignore-unfixed) without failing the build; the Security tab surfaces them.
2026-06-01 17:31:29 +02:00
3 changed files with 334 additions and 0 deletions

38
.github/workflows/detect-duplicate.yml vendored Normal file
View File

@@ -0,0 +1,38 @@
name: 🔁 Detect Duplicate Issues
on:
issues:
types: [opened]
permissions:
contents: read
concurrency:
group: detect-duplicate-${{ github.event.issue.number }}
cancel-in-progress: true
jobs:
detect:
name: 🔍 Find similar issues
if: github.actor != 'github-actions[bot]'
runs-on: ubuntu-24.04
permissions:
issues: write
contents: read
steps:
- name: 📥 Checkout repository
uses: actions/checkout@de0fac2e4500dabe0009e67214ff5f5447ce83dd # v6.0.2
- name: 🍞 Setup Bun
uses: oven-sh/setup-bun@0c5077e51419868618aeaa5fe8019c62421857d6 # v2.2.0
with:
bun-version: latest
- name: 🔍 Detect duplicate issues
run: bun scripts/detect-duplicate-issue.mjs
env:
GH_TOKEN: ${{ github.token }}
GITHUB_REPOSITORY: ${{ github.repository }}
ISSUE_NUMBER: ${{ github.event.issue.number }}
ISSUE_TITLE: ${{ github.event.issue.title }}
ISSUE_BODY: ${{ github.event.issue.body }}

60
.github/workflows/trivy-scan.yml vendored Normal file
View File

@@ -0,0 +1,60 @@
name: 🛡️ Trivy Security Scan
# Filesystem scan (Streamyfin ships no container image): finds vulnerable dependencies,
# leaked secrets and misconfigurations, and reports them to GitHub code scanning.
# Runs post-merge + weekly (not on PRs — dependency-review already gates PRs, and SARIF
# upload needs a write token that fork PRs don't get).
on:
push:
branches: [develop, master]
schedule:
- cron: "50 7 * * 5" # Weekly, Friday 07:50 UTC
workflow_dispatch:
permissions:
contents: read
concurrency:
group: trivy-${{ github.ref }}
cancel-in-progress: true
jobs:
trivy:
name: 🔎 Filesystem scan
runs-on: ubuntu-24.04
permissions:
contents: read
security-events: write # upload SARIF to code scanning
steps:
- name: 📥 Checkout repository
uses: actions/checkout@de0fac2e4500dabe0009e67214ff5f5447ce83dd # v6.0.2
# Rotate the DB cache weekly (matches the scheduled scan): cache hits within the week
# instead of a fresh immutable entry per run, still refreshing the DB every week.
- name: 🗓️ Compute weekly Trivy cache key
id: trivy-cache-key
run: echo "value=trivy-db-${{ runner.os }}-$(date -u +%G-%V)" >> "$GITHUB_OUTPUT"
- name: 💾 Cache Trivy vulnerability DB
uses: actions/cache@27d5ce7f107fe9357f9df03efb73ab90386fccae # v5.0.5
with:
path: ~/.cache/trivy
key: ${{ steps.trivy-cache-key.outputs.value }}
restore-keys: trivy-db-${{ runner.os }}-
- name: 🔎 Run Trivy filesystem scan
uses: aquasecurity/trivy-action@ed142fd0673e97e23eac54620cfb913e5ce36c25 # v0.36.0
with:
scan-type: fs
scan-ref: .
scanners: vuln,secret,misconfig
ignore-unfixed: true
severity: CRITICAL,HIGH
format: sarif
output: trivy-results.sarif
- name: 📤 Upload results to code scanning
uses: github/codeql-action/upload-sarif@7211b7c8077ea37d8641b6271f6a365a22a5fbfa # v4.36.0
with:
sarif_file: trivy-results.sarif
category: trivy-fs

View File

@@ -0,0 +1,236 @@
#!/usr/bin/env bun
/**
* Flags likely-duplicate issues when a new issue is opened, using lexical similarity
* (Jaccard over word sets of the title and body) — no API key, no embeddings.
*
* On a match it posts ONE comment listing the closest open issues and adds the
* "possible duplicate" label. If nothing is similar enough, it does nothing.
*
* Env:
* GITHUB_REPOSITORY owner/repo
* ISSUE_NUMBER the new issue number
* ISSUE_TITLE the new issue title
* ISSUE_BODY the new issue body
* GH_TOKEN/GITHUB_TOKEN for gh (provided in CI)
* DUP_THRESHOLD similarity threshold 0..1 (default 0.3)
* DUP_MAX max matches to report (default 5)
* DUP_FIXTURE optional path to a JSON array of {number,title,body} (local testing)
* DRY_RUN if set, print results instead of commenting/labelling
*/
import { execFileSync } from "node:child_process";
import { readFileSync } from "node:fs";
// Parse a numeric env var, falling back to `def` only when unset/empty/NaN so an explicit 0 is honoured.
const numEnv = (name, def) => {
const raw = process.env[name];
if (raw === undefined || raw === "") return def;
const n = Number(raw);
return Number.isNaN(n) ? def : n;
};
const REPO = process.env.GITHUB_REPOSITORY || "streamyfin/streamyfin";
const NUMBER = numEnv("ISSUE_NUMBER", Number.NaN);
const TITLE = process.env.ISSUE_TITLE || "";
const BODY = process.env.ISSUE_BODY || "";
const THRESHOLD = numEnv("DUP_THRESHOLD", 0.3);
const MAX = numEnv("DUP_MAX", 5);
const DRY = !!process.env.DRY_RUN;
const LABEL = "possible duplicate";
const MARKER = "<!-- duplicate-detector -->";
// Generic stop words only — keep domain/feature/platform words (android, downloads,
// subtitles…) since those are exactly what makes two reports the same or different.
const STOP = new Set(
(
"a an the and or but if then of to in on at by for with from as is are was were be been being do does did " +
"it its this that these those i you we they me my your our their he she him her " +
"when while where what which who how why so just then than too very can could would should will " +
"not no nor only own same s t don dont im ive please thanks hi hello also still get got use used using " +
"app application streamyfin issue bug"
).split(/\s+/),
);
const stem = (w) => w.replace(/(ing|ed|es|s)$/, "");
const tokens = (s) =>
(s || "")
.toLowerCase()
.replace(/```[\s\S]*?```/g, " ") // drop code blocks
.replace(/<!--[\s\S]*?-->/g, " ") // drop html comments
.replace(/https?:\/\/\S+/g, " ") // drop urls
.replace(/[^a-z0-9\s]/g, " ")
.split(/\s+/)
.filter((w) => w.length > 2 && !STOP.has(w))
.map(stem)
.filter((w) => w.length > 2);
const jaccard = (a, b) => {
const A = new Set(a);
const B = new Set(b);
if (!A.size || !B.size) return 0;
let inter = 0;
for (const x of A) if (B.has(x)) inter++;
return inter / (A.size + B.size - inter);
};
const newTitle = tokens(TITLE);
const newBody = tokens(BODY);
const score = (o) =>
0.6 * jaccard(newTitle, tokens(o.title)) +
0.4 * jaccard(newBody, tokens(o.body));
// fetch open issues (excluding PRs and the new issue itself)
let issues;
if (process.env.DUP_FIXTURE) {
issues = JSON.parse(readFileSync(process.env.DUP_FIXTURE, "utf8"));
} else {
const raw = execFileSync(
"gh",
[
"api",
`repos/${REPO}/issues`,
"--paginate",
"-X",
"GET",
"-f",
"state=open",
"-f",
"per_page=100",
"--jq",
".[] | select(.pull_request | not) | {number, title, body}",
],
{ encoding: "utf8", maxBuffer: 1e8 },
);
issues = raw
.split("\n")
.filter(Boolean)
.map((l) => JSON.parse(l));
}
const matches = issues
.filter((o) => o.number !== NUMBER)
.map((o) => ({ ...o, s: score(o) }))
.filter((o) => o.s >= THRESHOLD)
.sort((a, b) => b.s - a.s)
.slice(0, MAX);
if (!matches.length) {
console.log("No likely duplicates found.");
process.exit(0);
}
// Neutralise other issues' titles before echoing them back: break @mentions and
// strip markdown/HTML control chars so a maliciously-named issue can't ping people
// or inject formatting into our comment. GitHub linkifies "#123" on its own.
const safeTitle = (t) =>
(t || "")
.replace(/@/g, "@")
.replace(/[`<>|*_~[\]]/g, " ")
.replace(/\s+/g, " ")
.trim()
.slice(0, 140);
const list = matches
.map(
(m) =>
`- #${m.number}${safeTitle(m.title)} (≈ ${Math.round(m.s * 100)}% similar)`,
)
.join("\n");
const comment = [
MARKER,
"🔍 **This looks like it might be a duplicate.** Possibly related open issues:",
"",
list,
"",
"If yours is different, ignore this — a maintainer will confirm. Otherwise, please 👍 the existing issue and add any extra details there.",
].join("\n");
console.log(`Found ${matches.length} possible duplicate(s):\n${list}`);
if (DRY) {
console.log("\nDRY_RUN: not commenting/labelling.");
process.exit(0);
}
// Live mode needs a real issue number; refuse rather than POST to /issues/NaN/...
if (!Number.isInteger(NUMBER) || NUMBER <= 0) {
console.error(
`Invalid ISSUE_NUMBER ${JSON.stringify(process.env.ISSUE_NUMBER)} — refusing to comment.`,
);
process.exit(1);
}
// Idempotency: skip if we've already flagged this issue (guards re-runs / future triggers).
const priorComments = execFileSync(
"gh",
[
"api",
`repos/${REPO}/issues/${NUMBER}/comments`,
"--paginate",
"--jq",
".[].body",
],
{ encoding: "utf8", maxBuffer: 1e8 },
);
if (priorComments.includes(MARKER)) {
console.log("Already flagged (marker present); skipping.");
process.exit(0);
}
execFileSync(
"gh",
[
"api",
"-X",
"POST",
`repos/${REPO}/issues/${NUMBER}/comments`,
"-f",
`body=${comment}`,
],
{ stdio: "ignore" },
);
try {
execFileSync(
"gh",
[
"api",
"-X",
"POST",
`repos/${REPO}/issues/${NUMBER}/labels`,
"-f",
`labels[]=${LABEL}`,
],
{ stdio: "ignore" },
);
} catch {
// label may not exist yet — create then add
execFileSync(
"gh",
[
"api",
"-X",
"POST",
`repos/${REPO}/labels`,
"-f",
`name=${LABEL}`,
"-f",
"color=fbca04",
"-f",
"description=Automatically flagged as a possible duplicate",
],
{ stdio: "ignore" },
);
execFileSync(
"gh",
[
"api",
"-X",
"POST",
`repos/${REPO}/issues/${NUMBER}/labels`,
"-f",
`labels[]=${LABEL}`,
],
{ stdio: "ignore" },
);
}
console.log("Commented and labelled.");