Vendor: Cyan Eclipse Technologies · Platforms: MCP over stdio (HTTP transport on roadmap)
Certification scope
CERTIFIED(local) — Eclipse Bazaar Test Bench 2026-10-25, 21 pass, 1 na of 22 checks (U1 auth: stdio process boundary). Run by W-003 (Verity), suites/ledger.py. Run 1 FAILED card retained (bench-suite probe-ordering defect; product behavior was correct); run 2 is the standing card. Verity adjudication A1 on the card: the settled-backfill exclusion boundary matches spec section 3.3 intent.
What it does
- Limits engine with the industry's asymmetry built in: decreases apply immediately; increases wait out the stricter of the suite floor or the state's delay, then need a fresh confirmation or they lapse — no tool path exists that raises a limit, and a weakened policy file refuses to load
- Self-exclusion absolute and honored identically from any source: wagering surfaces deny EXCLUDED with the resource block attached; bookkeeping survives (settled pre-exclusion bets stay recordable — bench-adjudicated boundary); no early-end path exists
- Versioned jurisdiction matrix in which sportsbook, DFS-classic, and pick'em variants are distinct legal objects; staleness degrades verdicts to UNKNOWN-and-deny
- Unified ledger of the user's own bet records with honest cash P&L: bonus-bet stake is never cash (unsettled bonus valued at zero, stake-not-returned stated); headline P&L independently recomputed on the bench; pattern flags computed on ledger data only
- Fixture odds comparison ordered by price only, operator promotions stripped by construction; arb observations carry the books-limit-winning-accounts warning and execution_support: null
- Not a sportsbook. No bets are placed here. No placement tool, no account connection, and no credential is accepted anywhere — records enter only from the user's own exports and entries
Tools
profile_setprofile_getlimit_setlimit_confirmlimits_listtimeout_startexclusion_startexclusion_endexclusion_statusjurisdiction_checkmatrix_statusbet_record_addrecords_listdeposit_record_addpnl_summaryreality_snapshotodds_comparearb_observeslip_prepareadvice_requestresources_getreceipt_list
Problems it addresses
- My betting apps show 'profit' that counts free bets as cash
- I set a limit to protect myself and the app let me raise it in the same session
- I am self-excluded at one book and the others never got the memo
Not tested
- Any live odds feed — B0 odds are fixtures shaped like The Odds API v4 responses; no feed account exists and none was contacted (licensed signup is an owner step).
- Live-user operation of any kind — the spec's counsel gate (D4) precedes any live-user launch; B0 runs on synthetic records only.
- Real operator limit/exclusion sync — operator and state programs do not publish a consumer sync route; B0 honors user-declared + user-provided records identically (spec 3.3).
- Jurisdiction matrix freshness against live sources — B0 rows are fixture data as of 2026-10-25 with a 30-day TTL that degrades verdicts to UNKNOWN; a sourced refresh pipeline is a live-phase item.
- Placement (T2/T3) — no placement capability exists in v1 by design (spec 6); there is nothing to test because there is no such tool.
Bench scorecard: test-bench/results/eclipse-ledger-mcp-2026-10-25-2.json
Evaluation-intake listing
This listing records the tool's place in the Bazaar's evaluation
and monitoring pipeline, with the evidence gathered so far printed above — verdict, scope,
and not-tested items included. It is not sold on this site, and its presence in the catalog
is not a claim beyond that evidence. The service behind every verdict is the bench:
what a bench run buys →