Eclipse Namesake

Bench verdict: CERTIFIED(local, fixture registry snapshot)

Certified scope: CERTIFIED(local, fixture registry snapshot) — Verity bench of v0.1.0 (2026-10-27, argus101-program-20261027/wave2/bench-namesake.md) PASS-CONDITIONAL(local, fixture registry snapshot) with sole condition C1 (Finding 1, stored-read paths served unverified stored verdicts), plus delta re-bench of v0.1.1 in the same record (stored-read paths only): C1 CLOSED at scope (local, fixture registry snapshot) — no new findings; Finding 2 (snapshot envelope without snapshot_ref accepted) also CLOSED. Builder suite 33/33 at v0.1.0 and 37/37 at v0.1.1 (33 prior + 4 new TestStoredReadVerification C1 regressions) re-run clean-state by the bencher both times, the v0.1.0 suite also under a socket-denied harness; decisive fixture F-N9 independently recomputed by the bencher with a scorer rewritten from the spec text (builder scorer NOT imported) at both benches — agent-harborlight name 0.628205 / purpose 0.6 / capability 0.666667 / composite 0.627436 → REVIEW, server top_k[0] identical, fresh-process stdio verify from artifacts alone verified:true / VERIFIED_EXPORT. Boundary pins exact at v0.1.0 bench: exact 'Payseal!' composite 0.500000 → COLLISION by exact-name override, near 'Payseel' 0.454286 → REVIEW, stem-only 'Ledger Lynx' 0.242424 → CLEAR. Delta probes ALL_C1_OK: intact control serves the sealed CLEAR verdict normally on every path (screen_list carries verified:true / VERIFIED_EXPORT); score-tamper (stored CLEAR→COLLISION, max 0.999999, hash/signature NOT recomputed) and Herald-signature-tamper (signature f*64) both downgrade on verdict_get, breakdown_get, screen_list (incl. verdict-filtered lists: COLLISION=0, CLEAR=0) and idempotent re-screen to UNVERIFIED_EXPORT / verdict_claim null / NAMESAKE_VERIFY_FAIL:<step>, metering unchanged; a snapshot envelope without snapshot_ref now refuses SNAPSHOT_INCOMPLETE with 0 screens / 0 metering, and verify over a ref-less snapshot downgrades at step snapshot_hash. Scope is exactly: local stdio MCP, fixture registry snapshot only, scoring version namesake-score/1.0 (weights name 0.5 / purpose 0.3 / capability 0.2; thresholds REVIEW ≥ 0.40, COLLISION ≥ 0.70 plus exact-normalised-name override), value_class 'test' throughout — now including the stored-read paths (namesake_verdict_get, namesake_breakdown_get, namesake_screen_list, and the idempotent path of namesake_screen), which recompute-verify the stored screen before serving anything; no live registry is read, nothing is registered, reserved, assigned, blocked, renamed, or revoked, and no counsel-gate (SPEC §10) live phase is built or claimed. Novel mechanism (SPEC §2.1, verbatim): similarity scoring over registered agent identities (name, purpose, capability profile) issues a collision verdict before an agent launches under a confusing identity. Non-blocking bench observation Finding 3 (dispatch does not enforce schema additionalProperties:false — unexpected extra args are silently ignored, with no state or verdict effect observed) remains open, untouched. A Namesake verdict is a similarity screen over a stated snapshot — it is not a trademark opinion, not a legal clearance, and not permission to launch (SPEC §11.2).

Not tested:

• Live-registry screening, any verdict carrying consequences in a real registry, or any real agent identity — fixture registry snapshot only; the SPEC §10 counsel gate is unbuilt by design, named, with no dormant live mode: screening against a live registry, a verdict carrying consequences in a real registry (blocking, flagging, or delaying a real launch), or any public marketing of the screen as clearance — each is a separate, counsel-cleared, re-benched build, with counsel review of (a) trademark-adjacent claim language and (b) the consequences a verdict may carry
• Any trademark opinion, legal clearance, or permission to launch — a Namesake verdict is a similarity screen over a stated snapshot (SPEC §11.2): a CLEAR verdict does not mean a name is legally available, non-infringing, or registrable anywhere, and a COLLISION verdict does not mean a name infringes anything; a CLEAR verdict means only 'no collision above threshold in this snapshot under namesake-score/1.0'. The SPEC §11.4 forbidden claim strings are never claimed for v0.1.1 (they are listed in SPEC §11.4 only and are not reproduced on this surface); no wording implies live registry coverage, registration, or legal advice, and no wording implies a verdict outcome can be bought (SPEC §9 / L7: metering attaches to the screen, identical for CLEAR, REVIEW, and COLLISION). Cross-script and homoglyph collisions are a known blind spot of v0.1 scoring (SPEC Q-N3): the fixture normaliser covers the Latin-script fixture set only, and no listing implies coverage the scorer does not have
• Production signing / public-key verification — v0.1.1 signing is a fixture HMAC stand-in for the Herald pattern (local fixture keys, keys/ gitignored, mode 600, never committed); the live RS256 Herald signing form is unbuilt and counsel-gated. Verification uses pinned fixture keys only — keys are pinned out-of-band and never read from, or selected by, a record under verification; the pinned key file holds shared HMAC secrets, not public keys, and no public-key or third-party verifiability is claimed. A fresh clone generates its own fixture keys to run the full suite
• Hosted or live operation — local stdio only; metering points only per SPEC §9 (exactly one Meter record per completed namesake_screen reaching a sealed verdict; refusals and same-snapshot idempotent re-screens emit none; CLEAR, REVIEW, and COLLISION meter identically), the $5/screen shape is a book reference with no charge path built and no live Meter ingest

Live state note: Bench evidence: argus101-program-20261027/wave2/bench-namesake.md (Verity bench of v0.1.0, 2026-10-27, PASS-CONDITIONAL(local, fixture registry snapshot), sole condition C1/Finding 1; builder suite re-run clean-state 33/33 twice plus a socket-denied run, determinism byte-identical across fresh stores, F-N9 decisive fixture independently recomputed with a scorer rewritten from the spec text, adversarial pass: forged builder signature, forged Herald signature, key-supplied-inside-record, full forgery with recomputed record_hash, forged Genesis entry signature, presented-score/verdict substitution battery — all named refusals or §6.3 downgrades with verdict_claim null; snapshot entry removal/byte-drift refused; malformed-args battery on every tool on both surfaces with no tracebacks; refusal-must-not-mutate verified; key custody 600/gitignored/values absent from source; metering exactness reproduced (3 sealed screens = 3 points, 0 for refusals and idempotent re-screens, value_class 'test' on 100% of artifacts checked); scope audit §10 unbuilt with fence startup refusals bench-reproduced; surface = exactly 4 builder + 3 operator tools disjoint, no registry-write verbs; §11.4 forbidden claim strings absent tree-wide excl. SPEC.md's own forbidding enumeration) + delta re-bench of v0.1.1 in the same record (suite 37/37 clean-state twice, bencher's own on-disk store-tamper C1 probes ALL_C1_OK for score-tamper and signature-tamper across all four stored-read paths, snapshot_ref strictness closed, F-N9 regression F_N9_MATCH=True; C1 CLOSED, Finding 2 CLOSED, no findings in delta scope). Fix record: argus101-program-20261027/wave2/fix-namesake.md. Independent scoring reproduction at v0.1.0 bench: composite 0.627436 → REVIEW for the F-N9 candidate, identical to the server. No test-bench scorecard was issued for this verdict and none is claimed (scorecard_ref/card_url null).