io.github.PatrickPi1312/eu-compliance

What we checked

Six static checks, weighted by risk. Every result reflects only what could be observed in the published package and repository — never intent.

• Injection surface

  pass25% of grade

  Tool descriptions/manifest scanned for instruction-injection patterns (imperatives at the model, hidden text, 'ignore previous', data-exfil URLs).

  Scanned 264 text field(s); no instruction-injection patterns found. PASS = no static red flag. A server can serve different tool descriptions at runtime (the rug-pull class), which static scanning cannot see.

• Supply chain

  pass25% of grade

  Package provenance: namespace verification, repo linkage, maintainer count, account age, postinstall scripts, typosquat distance.

  repository linkage verified

• Credential hygiene

  fail15% of grade

  How the server takes secrets (env vs plaintext config vs hardcoded); secrets appearing in tool schemas.

  tool "us_import_readiness_guard" has a secret-shaped input parameter "recovery_secret" — secrets must never be tool parameters (the model would supply/handle the credential). tool "recover_us_import_readiness" has a secret-shaped input parameter "recovery_secret" — secrets must never be tool parameters (the model would supply/handle the credential).

• Permission scope

  unscannable15% of grade

  Declared tools vs. breadth (filesystem, network, exec); flags shell-exec and unbounded filesystem access.

  No source files, package.json, or pyproject were fetched; cannot infer capability scope.

• Version behavior

  warn10% of grade

  Diff of tool definitions between versions; new permissions or changed descriptions in a patch release (the postmark-mcp class).

  vs prior 1.1.0 (minor bump): tool "validate_vat" description changed (wording only, no injection pattern); tool "validate_iban" description changed (wording only, no injection pattern); tool "prepare_agentllm_micro" description changed (wording only, no injection pattern); tool "agentllm_micro" description changed (wording only, no injection pattern); tool "prepare_invoice_to_pay_eu" description changed (wording only, no injection pattern); tool "invoice_to_pay_dossier_eu" description changed (wording only, no injection pattern); tool "prepare_us_import_readiness" description changed (wording only, no injection pattern); tool "us_import_readiness_guard" description changed (wording only, no injection pattern); tool "recover_us_import_readiness" description changed (wording only, no injection pattern); tool "screen_sanctions_eu" description changed (wording only, no injection pattern); tool "eu_vat_rules" description changed (wording only, no injection pattern); tool "eu_vat_rules" gained input property(ies): customer_vat_valid, above_oss_threshold; tool "check_counterparty_eu" description changed (wording only, no injection pattern); tool "check_counterparty_eu" gained input property(ies): own_country, relationship, type, b2b; tool "validate_einvoice_eu" description changed (wording only, no injection pattern); tool "verify_compliance_receipt" description changed (wording only, no injection pattern); tool "must_verify_before_pay" description changed (wording only, no injection pattern); tool "lookup_company_eu" description changed (wording only, no injection pattern); tool "agent_spend_statement" description changed (wording only, no injection pattern); tool "market_data" description changed (wording only, no injection pattern); tool "market_history" description changed (wording only, no injection pattern); tool "ai_act_disclosure" description changed (wording only, no injection pattern); tool "tx_preflight" description changed (wording only, no injection pattern); tool "tx_status" description changed (wording only, no injection pattern); tool "token_status" description changed (wording only, no injection pattern); tool "read_url" description changed (wording only, no injection pattern); new tool "recommend_agent_tool" touches a sensitive surface (email/exec/net/fs); new tool "prepare_agentllm" appeared; new tool "agentllm_basic" touches a sensitive surface (email/exec/net/fs); new tool "agentllm_plus" touches a sensitive surface (email/exec/net/fs); new tool "rpc_read_batch" touches a sensitive surface (email/exec/net/fs); new tool "wallet_balance" appeared; new tool "gas_estimate" appeared; new tool "contract_info" appeared; new tool "spot_price" appeared; new tool "open_social_search" appeared; new tool "perp_market_data" appeared; new tool "coin_positioning_brief" appeared; new tool "market_decision_pack" appeared; new tool "soft_commodity_stress" appeared; new tool "prepare_x402_purchase_guard" appeared; new tool "x402_purchase_guard" appeared.

• Transport config

  fail10% of grade

  Remote servers: TLS and auth mode (none/token/OAuth). Local servers: whether the manifest indicates it phones home.

  remote MCP endpoint declares NO authentication (unauthenticated remote — nginx-ui/CVE-2026-33032 class)