French Business Analyser

v2.2.0

Verified French business data for autonomous B2B agents. Without this MCP, agents hallucinate financial data. With it, they get real-time signals from 9 offi...

1· 47·0 current·0 all-time
MIT-0
Download zip
LicenseMIT-0 · Free to use, modify, and redistribute. No attribution required.
Security Scan
Capability signals
CryptoRequires walletCan make purchasesRequires OAuth token
These labels describe what authority the skill may exercise. They are separate from suspicious or malicious moderation verdicts.
VirusTotalVirusTotal
Benign
View report →
OpenClawOpenClaw
Benign
medium confidence
Purpose & Capability
Name/description match the actions described in SKILL.md: querying French registries, synthesising results with Claude, and offering invoice parsing and scoring. There are no unexpected agent-side binaries or credentials required; server-side env vars (Anthropic key, INSEE key, Redis, Stripe, x402 parameters) are consistent with running such a service.
Instruction Scope
The SKILL.md explicitly instructs agents to call a hosted MCP endpoint (Railway URL) for lookups and to send SIREN/VAT numbers, optional IBAN hashes, and optionally invoice text to the service. It defines a consent gate for paid tools and forbids calling certain endpoints without user approval. This is appropriate for the problem domain but means confidential business data will traverse a third-party server; the skill's correctness depends on the agent following the consent rules (the agent enforces them).
Install Mechanism
Instruction-only skill with no install spec and no code bundled to execute on the agent — lowest-risk installation footprint. Self-host instructions point to a public GitHub repo and a standard Docker workflow, which is expected.
Credentials
Agent-side requires no env vars (proportionate). The _meta.json lists server-side env vars (INSEE_API_KEY, ANTHROPIC_API_KEY, REDIS_URL, STRIPE and x402 payment config, METRICS_SECRET_TOKEN) which are reasonable for a hosted lookup+payment service. Ensure you only provide those to a trusted self-host or to operators you trust; no agent-side secret is required.
Persistence & Privilege
Skill is not always-enabled and is user-invocable; it does not request persistent agent privileges. It relies on an external hosted service (default) which keeps short TTL caches per the docs; persistence concerns relate to that remote server, not this instruction-only skill.
Assessment
What to consider before installing: - Privacy: the hosted endpoint will receive SIREN/VAT numbers and (if you use it) invoice text and IBANs. For confidential invoices or sensitive payment identifiers, prefer self-hosting and audit the repository before sending data to the hosted service. - Consent gate: the SKILL.md relies on the agent to block paid-tool calls until the user consents. Verify your agent enforces that consent flow; otherwise you could incur charges or leak data. - Verify the hosted endpoint and source: the README points to a Railway URL and a GitHub repo. Confirm the repo is the exact source of the hosted service (watch for forks or typosquatting) and review .env.example to see what server operators must provide. - Third parties: the server uses Anthropic for synthesis and may contact official registries; server operators will hold Anthropic and registry API keys. Understand their privacy policy and how they handle traveler data (logs, telemetry). - Payments: the service supports x402 (USDC on Base) and Stripe MPP; confirm billing behavior and that no wallet private key is required on the agent side (the docs claim verification-only). - If you lack resources or trust: self-host via the provided Docker instructions, supply your own INSEE/Anthropic keys, and run the service in your environment. That gives the strongest control and makes the skill's claims verifiable. Overall: the skill appears coherent for its stated purpose, but the main risk is that sensitive business documents/identifiers will be sent to a third-party hosted server. If you plan to process confidential invoices or payment info, audit or self-host before use.

Like a lobster shell, security has layers — review code before you run it.

b2bvk9791kc4n3xv9gfkt7hezggxex846qsgdue-diligencevk9791kc4n3xv9gfkt7hezggxex846qsgfrancevk9791kc4n3xv9gfkt7hezggxex846qsginvoicevk9791kc4n3xv9gfkt7hezggxex846qsglatestvk97c43xpm84a7m80dnbse6pkj5849ykmmcpvk9791kc4n3xv9gfkt7hezggxex846qsgsuppliervk9791kc4n3xv9gfkt7hezggxex846qsg

License

MIT-0
Free to use, modify, and redistribute. No attribution required.

Comments