skills: {'id': 'cross_lens_verify', 'name': 'cross_lens_verify', 'description': "A2 — the cross-lens join. Fuse TunnelMind's two lenses (Scry attacker\nintelligence + Sigil supply graph) into ONE verdict on a single node key.\nThis is the moat: no siloed competitor owns both halves of the graph, so\nthe fused `cross_lens` block carries information neither lens can supply\nalone.\n\nUse this tool when:\n- An agent must decide whether to transact with an IP, domain, ASN, or\n entity_slug, and a one-lens answer is not enough.\n- You want a single composite trust verdict instead of running\n Scry + Sigil calls separately and reconciling them by hand.\n\nInputs:\n- `node` (required): an IPv4 address, a domain, an ASN (e.g. `AS64500`),\n or an entity_slug. Type is auto-detected.\n- `weights` (optional): per-component weight overrides.\n- `thresholds` (optional): `{ pass, fail }` verdict cutoffs (defaults 0.7 / 0.3).\n- `ait` (optional): an ATAP AIT id. When present, the verdict is chained\n onto the AIT as a witness-tier `cross_lens:verified` event signed by\n Sigil (witness OAI-2026-0000201) — replayable evidence, not just JSON.\n\nReturns: per-lens `scry` + `sigil` blocks (transparency), a fused\n`cross_lens` block with `verdict` / `trust_score` / `confidence` /\n`signals` / `recommendations`, a 5-minute signed `sigil_token`, and a\n`witnessed_event` block when an AIT was supplied.\n\nFailure semantics: each lens fails independently. Single-lens answers\nstill return 200 with a `confidence` of 0.55. Returns 503 only when BOTH\nlenses are unavailable.", 'tags': [], 'examples': None, 'input_modes': None, 'output_modes': None}, {'id': 'sigil_verify_supply_path', 'name': 'sigil_verify_supply_path', 'description': 'The core pre-bid check. Verify the trustworthiness of one programmatic ad\nsupply path and get back a composite trust verdict plus a signed proof token.\nSigil composes ads.txt authorization, datacenter-IP classification, Scry\nfraud-corpus lookup, and app-bundle checks into one score.\n\nUse this tool when:\n- An ad-buying agent is about to bid and must confirm the supply is genuine.\n- You want one call instead of running ads.txt / IP / bundle checks separately.\n\nInputs:\n- `supply_path` (required): { publisher_domain, exchange, seller_id, and\n optionally ip_address, app_bundle:{bundle_id,platform} }.\n- `ait` (optional): an ATAP AIT id — when present, Sigil records this\n verification as a witnessed attestation event and binds the token to it.\n\nReturns: `trust_score` (0-1), `verdict` (pass/warn/fail/unknown), per-check\nresults, `recommendations`, and a signed `sigil_token` (5-min) to attach to\nthe bid as proof. The submitted IP is used for lookup only — never stored.', 'tags': [], 'examples': None, 'input_modes': None, 'output_modes': None}, {'id': 'sigil_verify_ads_txt', 'name': 'sigil_verify_ads_txt', 'description': "Check whether an exchange/SSP is authorized to sell a publisher's inventory,\nper the publisher's ads.txt file. Fast cached lookup against Sigil's daily\ncrawl of the top ~10k publisher domains.\n\nUse this tool when:\n- You need a single, narrow authorization check (not a full supply-path score).\n- You are validating a (publisher, exchange, seller_id) triple from a bid request.\n\nInputs:\n- `publisher_domain`, `exchange_domain`, `seller_id` (all required).\n- `resolve_chain` (optional): when true and the entry is RESELLER, Sigil walks\n one hop into the exchange's sellers.json to identify the upstream seller.\n\nReturns: `verified` (true/false/null), `confidence`, the matched ads.txt entry,\nand any `warnings` (e.g. seller_type mismatch).", 'tags': [], 'examples': None, 'input_modes': None, 'output_modes': None}, {'id': 'sigil_verify_ip_type', 'name': 'sigil_verify_ip_type', 'description': 'Classify an IPv4 address as datacenter, residential, mobile, or unknown.\nDetects datacenter traffic posing as real user devices. Stateless — the IP\nis never logged or stored.\n\nUse this tool when:\n- You need to know whether bid-request traffic originates from a datacenter.\n\nInputs:\n- `ip` (required): an IPv4 address.\n\nReturns: `ip_type`, `confidence` (high/medium/low), and the ASN + AS-org name.', 'tags': [], 'examples': None, 'input_modes': None, 'output_modes': None}, {'id': 'sigil_verify_app_bundle', 'name': 'sigil_verify_app_bundle', 'description': 'Verify that a mobile/CTV app bundle ID actually exists in its app store and,\noptionally, that the listed developer matches. Detects bundle-ID spoofing in\nbid requests.\n\nUse this tool when:\n- A bid request names an app bundle and you must confirm the app is real.\n\nInputs:\n- `bundle_id` (required), `platform` (required: ios | android | ctv_* | web),\n- `claimed_developer` (optional): developer name to match against the listing.\n\nReturns: `verified` (true/false/null), the store listing, and `developer_match`.', 'tags': [], 'examples': None, 'input_modes': None, 'output_modes': None}, {'id': 'sigil_verify_supply_chain', 'name': 'sigil_verify_supply_chain', 'description': 'Verify a full OpenRTB SupplyChain (schain) object — every node, end to end.\nPer node Sigil checks the seller against the exchange sellers.json and the\norigin ads.txt, then returns a per-node and aggregate verdict plus a signed\ntoken.\n\nUse this tool when:\n- A bid request carries an OpenRTB `schain` and you want it verified verbatim.\n\nInputs:\n- `schain` (required): an OpenRTB SupplyChain object ({ ver, complete,\n nodes:[{asi,sid,hp}] }).\n- `site_domain` or `app_bundle` (optional): the inventory origin, checked\n against node[0] via ads.txt / OWNERDOMAIN.\n\nReturns: per-node `nodes` results, an aggregate `verdict`, `recommendations`,\nand a signed `sigil_token`.', 'tags': [], 'examples': None, 'input_modes': None, 'output_modes': None}, {'id': 'traverse_supply_chain', 'name': 'traverse_supply_chain', 'description': "Walk the supply graph for a publisher domain and get back the ITEMIZED\nsell paths — distinct from sigil_verify_supply_chain (which verifies a\nschain you BRING) and from the dark-pool-risk signal (which only returns\ncounts). Here Sigil reconstructs the paths from its own crawl: every SSP\nthe publisher declares it sells through, joined to that SSP's identity and\nclassified two-sided against the SSP's sellers.json.\n\nUse this tool when:\n- You have a publisher domain but no schain, and want to SEE its real\n authorized supply paths and where the opacity is.\n- dark-pool-risk flagged a publisher and you need the specific contradicted\n paths driving the risk, not just the aggregate.\n\nInputs:\n- `domain` (required): the publisher domain, e.g. `cnn.com`.\n- `limit` (optional): max paths returned (default 200, cap 500). The list\n is ordered riskiest-first (contradicted, then reseller) so a truncated\n page is still the most useful; the `supply_paths` counts are always over\n the FULL set.\n\nReturns: `supply_paths` aggregate counts (total / direct / reseller /\ncorroborated / contradicted / unchecked) and `paths[]`, each with the SSP\nidentity, `seller_id`, `seller_type`, `klass` (corroborated = seat present;\ncontradicted = SSP crawled but seller_id absent → real risk; unchecked =\nSSP not yet crawled → not risk), and `resells_to` (one level of downstream\nreseller expansion). Returns in_supply_graph:false if the domain is not in\nthe crawled corpus.", 'tags': [], 'examples': None, 'input_modes': None, 'output_modes': None}, {'id': 'sigil_score_entity', 'name': 'sigil_score_entity', 'description': 'Get the pre-computed trust score for one supply-chain entity (a publisher or\nan SSP). Scores are recomputed daily from ads.txt health, supply-chain\ndirectness, reach, and stability — deterministic, no ML black box.\n\nUse this tool when:\n- You want a fast standing trust signal for an entity without running checks.\n\nInputs:\n- `entity_id` (required): `{type}:{domain}` — e.g. `publisher:nytimes.com` or\n `ssp:pubmatic.com`.\n\nReturns: `trust_score` (0-1), `score_components`, the 14-day `trend`, and\n`warnings`.', 'tags': [], 'examples': None, 'input_modes': None, 'output_modes': None}; uptime_30d 1.0%; p95 263.0ms; conformance: pass
How to connect
https://mcp.sigil.tunnelmind.ai/mcpcurl -X POST https://mcp.sigil.tunnelmind.ai/mcp \
-H 'Content-Type: application/json' \
-H 'Accept: application/json, text/event-stream' \
-d '{"jsonrpc":"2.0","id":1,"method":"initialize","params":{}}'