Wire Hlido into your code, your CI, and your agents.

@hlido/cli — score any agent from your terminal.

One command, no install, no sign-up. Pulls the latest published scorecard from hlido.eu and renders an inline summary. Good for CI scripts, ad-hoc spot checks, and pre-purchase due diligence.

# Inline scorecard for any reviewed agent
npx @hlido/cli check cursor

# Search the corpus
npx @hlido/cli search "code review"

# Side-by-side comparison
npx @hlido/cli compare cursor aider claude-code

# Top-tier list (VITAL only)
npx @hlido/cli tier vital --category coding

hlido-gate — block PRs that drop your score.

A composite action that fetches your latest signed Trust Attestation, posts (or refreshes) a sticky PR comment with score / tier / last-tested, and — in fail mode — fails the check below a threshold you choose. Vendors use this to keep public claims honest as the codebase changes.

# .github/workflows/hlido-gate.yml
name: Hlido Gate
on:
  pull_request:
    types: [opened, synchronize, reopened]

permissions:
  pull-requests: write
  contents: read

jobs:
  hlido-gate:
    runs-on: ubuntu-latest
    steps:
      - uses: ankitkapur1992-hlido/hlido-public/actions/hlido-gate@main
        with:
          slug: cursor          # your Hlido slug
          min-score: 70
          mode: warn            # flip to 'fail' once trusted
          comment-pr: "true"

Plug Hlido into Claude, Cursor, Cline, Continue, Goose.

JSON-RPC 2.0 endpoint at https://hlido.eu/mcp. Twelve tools: trust check, full scorecard, semantic search, claim verification, side-by-side compare, request audit, report issue, submit agent, and the new v2 trio — subscribe (advisory), explain (verdict reasoning), and recommend (task → ranked agent list). Hit tools/list to enumerate.

# ~/.claude.json (or Cursor → Settings → MCP)
{
  "mcpServers": {
    "hlido": {
      "type": "http",
      "url": "https://hlido.eu/mcp"
    }
  }
}

# Ad-hoc smoke test
curl https://hlido.eu/mcp \
  -X POST -H 'Content-Type: application/json' \
  -d '{"jsonrpc":"2.0","id":1,"method":"tools/list"}'

Full tool reference → · Open the playground →

See the Hlido score on every agent vendor page.

Sideload-only MVP for Chrome, Firefox, and Edge. Detects when a tab is on a reviewed-agent vendor page and surfaces the Laddoo Score, tier, and last-tested date in a popup with a click-through to the full review.

Quick install — Chrome/Edge:

1. Clone git clone https://github.com/ankitkapur1992-hlido/hlido-public.git
2. Open chrome://extensions, enable Developer mode
3. Click Load unpacked and pick the extension/ folder
4. Visit any reviewed-agent vendor page — the popup populates automatically

Firefox is a near-identical flow via about:debugging → Load Temporary Add-on. Listing on the Chrome Web Store / AMO is queued; sideload works in the meantime.

Verifiable per-agent attestations for agent-to-agent trust.

Each scored agent has a static, signed JSON attestation with the Laddoo Score, tier, headline metrics, last-tested date, and the canonical scorecard URL. Signed with HMAC-SHA256 using a key controlled by Hlido — anyone with the public verification key can confirm a payload has not been tampered with.

curl -s https://hlido.eu/data/attestations/aider.json | jq

# {
#   "schema_version": "1.0",
#   "slug": "aider",
#   "score": 90,
#   "tier": "VITAL",
#   "last_tested": "2026-04-26",
#   "scorecard_url": "https://hlido.eu/data/scorecards/aider.json",
#   "signature": { "alg": "HMAC-SHA256", "value": "..." }
# }

An Ed25519 signed-payload variant is on the roadmap for fully trustless offline verification. The current HMAC scheme is sufficient for live API consumers that already trust the hlido.eu TLS chain.

The full corpus as JSONL.

Every published review with sanitized claim-vs-evidence detail, integrity checksum, manifest, and a permissive license. Built for academic citation, AI-training corpora, and downstream analysis. Methodology weights stay private; results are open.

curl -L -O https://hlido.eu/data/open/hlido-corpus.jsonl
curl -L -O https://hlido.eu/data/open/hlido-corpus.jsonl.sha256
sha256sum -c hlido-corpus.jsonl.sha256
# hlido-corpus.jsonl: OK

Open the data landing page → · HuggingFace dataset (CC-BY-NC) →

Score badge and live sparkline — paste an iframe.

Two free embeds for any reviewed agent. The score badge (one div + one script) renders the current Laddoo state with a click-through to the review. The sparkline shows score history as a small SVG widget — useful on changelog pages and vendor sites.

# Score badge (live state, paste on your site)
<div id="hlido-badge-aider"></div>
<script src="https://hlido.eu/embed/aider/badge.js" async></script>

# Score-history sparkline
<iframe src="https://hlido.eu/embed/sparkline/aider/"
        width="320" height="80" frameborder="0"></iframe>

Both embeds are free. The optional Verified Badge subscription (€19 / €79) adds operational support around badge state — it is not a paywall on the embed itself.

/hlido — independent agent reviews from inside Claude Code.

Drop the skill into your .claude/skills/ folder and trigger Hlido lookups inline: "is Cursor still VITAL?", "compare Aider and Goose", or "what does Hlido say about <agent>?" The skill calls the public registry, scorecards, and attestations directly — no key required.

Pull requests, integrations, and re-use.

The public mirror at github.com/ankitkapur1992-hlido/hlido-public holds the CLI, the GitHub Action, the browser extension, the Claude Code skill, and the sanitized corpus. PRs welcome for new clients, language bindings, integrations, and bug fixes. Issues with any developer surface go on the same repo.

Have an idea for an integration we should ship next? Email [email protected] with a one-paragraph note.

• Public GitHub mirrorCLI + Action + extension + skill
• Issue trackerBugs, feature requests, integration ideas
• Integrations hubDistribution surfaces + REST endpoints
• Platform changelogWhat we shipped, by date