cloud.dchub/datacenter-power-grid-fiber

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

  fail25% of grade

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

  tool "register_standing_intent" description: names an outbound copy/exfil destination (BCC/exfil verb). tool "get_hosting_capacity" param "utility" description: cross-tool redirection ('instead of call…'). tool "set_shortlist_alert" param "notify" description: off-vendor URL in a description (...); capability text should not name where data goes. tool "set_market_alert" description: off-vendor URL in a description (hooks.example.com); capability text should not name where data goes. tool "summarize_for_citation" description: off-vendor URL in a description (dchub.cloud); capability text should not name where data goes. tool "get_permitting_intel" description: off-vendor URL in a description (dchub.cloud); capability text should not name where data goes. tool "register_standing_intent" description: off-vendor URL in a description (hooks.example.com); capability text should not name where data goes. resource "llms-txt" description: off-vendor URL in a description (dchub.cloud); capability text should not name where data goes. resource "manifest" description: off-vendor URL in a description (dchub.cloud); capability text should not name where data goes. server instructions: imperative telling the model it must read/send/attach something. server instructions: off-vendor URL in a description (dchub.cloud); capability text should not name where data goes.

• Supply chain

  warn25% of grade

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

  install hook "prepare" present and not a recognized benign build command; dependency "jose" is edit-distance 2 from popular package "@babel/core" (typosquat candidate)

• Credential hygiene

  fail15% of grade

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

  tool "search_facilities" has a secret-shaped input parameter "mpp_credential" — secrets must never be tool parameters (the model would supply/handle the credential). tool "get_market_intel" has a secret-shaped input parameter "mpp_credential" — secrets must never be tool parameters (the model would supply/handle the credential). tool "get_interconnection_queue" has a secret-shaped input parameter "mpp_credential" — secrets must never be tool parameters (the model would supply/handle the credential). tool "get_grid_data" has a secret-shaped input parameter "mpp_credential" — secrets must never be tool parameters (the model would supply/handle the credential). tool "analyze_site" has a secret-shaped input parameter "mpp_credential" — secrets must never be tool parameters (the model would supply/handle the credential). tool "compare_sites" has a secret-shaped input parameter "mpp_credential" — secrets must never be tool parameters (the model would supply/handle the credential). tool "get_fiber_intel" has a secret-shaped input parameter "mpp_credential" — secrets must never be tool parameters (the model would supply/handle the credential). tool "get_energy_prices" has a secret-shaped input parameter "mpp_credential" — secrets must never be tool parameters (the model would supply/handle the credential). tool "get_renewable_energy" has a secret-shaped input parameter "mpp_credential" — secrets must never be tool parameters (the model would supply/handle the credential). tool "get_water_risk" has a secret-shaped input parameter "mpp_credential" — secrets must never be tool parameters (the model would supply/handle the credential). tool "get_grid_intelligence" has a secret-shaped input parameter "mpp_credential" — secrets must never be tool parameters (the model would supply/handle the credential). tool "rank_markets" has a secret-shaped input parameter "mpp_credential" — secrets must never be tool parameters (the model would supply/handle the credential). tool "site_selection_canvas" has a secret-shaped input parameter "mpp_credential" — secrets must never be tool parameters (the model would supply/handle the credential). tool "bind_email" has a secret-shaped input parameter "api_key" — secrets must never be tool parameters (the model would supply/handle the credential).

• Permission scope

  pass15% of grade

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

  No shell-exec, unbounded/sensitive filesystem, or download-and-run primitives found in scanned source. PASS = no static red flag. Static scanning is blind to code fetched or executed at runtime.

• Version behavior

  fail10% of grade

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

  vs prior 2.12.8 (patch bump): tool "search_facilities" description changed (wording only, no injection pattern); tool "search_facilities" gained input property(ies): mpp_pay, mpp_credential; tool "semantic_search" description changed (wording only, no injection pattern); tool "get_interconnection_queue" gained input property(ies): mpp_pay, mpp_credential; tool "get_grid_data" gained input property(ies): mpp_pay, mpp_credential; tool "get_energy_prices" gained input property(ies): mpp_pay, mpp_credential; tool "get_renewable_energy" gained input property(ies): mpp_pay, mpp_credential; tool "get_water_risk" gained input property(ies): mpp_pay, mpp_credential; tool "why_dchub" description changed (wording only, no injection pattern); tool "rank_markets" gained input property(ies): mpp_pay, mpp_credential; tool "unlock_more_data" description changed (wording only, no injection pattern); new tool "find_sites" touches a sensitive surface (email/exec/net/fs); new tool "get_subsea_cables" touches a sensitive surface (email/exec/net/fs); new tool "get_peering_intel" touches a sensitive surface (email/exec/net/fs).

• Transport config

  pass10% of grade

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

  remote endpoint "https://dchub.cloud/mcp/officialregistry" uses TLS (https); remote endpoint uses OAuth (strongest auth mode)