Cite Holmes — Deep Research × Hallucination-Free Citations
Deep research with machine-verified citations
as observed 2026-11-22T06:34:16.981Z- Identifier
cite-holmes- Source
- ClawHub
- Version observed
- 6.9.0
- Source repository
- not published
- Repository observation
- No source repository listed
- First observed here
- 2026-08-28T22:54:40.332Z
- Observations recorded
- 40
- Installs (reported upstream)
- 5
- Weekly downloads (upstream)
- 2,899
- Declared license
- MIT-0
Observation history
Fields that differed: changelog latestVersion
| Field | Before | After |
|---|---|---|
changelog |
"v6.6.0: Consistency and transport hardening: license labeling unified as MIT across LICENSE/SKILL/README (removing the mismatch flagged in review); citation reachability probing n | "v6.9.0:审稿预演错误码驱动版。对抗审查问题单(--grill 审稿人预演)从按判定档模板升级为错误码驱动:12 类机器可读错误码各带一条针对性追问(与判定解释注册表同源,单一来源),历史报告的旧码自动经别名映射译为新码问题;问题单内同码去重;只出问题不改判定(自查工具非二次审判的边界不变)。761 项测试全绿,零新增依赖。" |
latestVersion |
"6.6.0" | "6.9.0" |
Fields that differed: changelog latestVersion
| Field | Before | After |
|---|---|---|
changelog |
"v6.5.0: Shipped verifiability and holdout scale-up: a manifest-driven sealed test subset now ships in-package (selfcheck --tests re-verifies verdict semantics and engineering lock | "v6.6.0: Consistency and transport hardening: license labeling unified as MIT across LICENSE/SKILL/README (removing the mismatch flagged in review); citation reachability probing n |
latestVersion |
"6.5.0" | "6.6.0" |
Fields that differed: changelog latestVersion
| Field | Before | After |
|---|---|---|
changelog |
"v6.4.0: Transport layer gains TLS-fingerprint downgrade retry: TLS-class handshake timeouts retry once with a TLS1.2 context (fixes an environment-level issue where some CDN-hoste | "v6.5.0: Shipped verifiability and holdout scale-up: a manifest-driven sealed test subset now ships in-package (selfcheck --tests re-verifies verdict semantics and engineering lock |
latestVersion |
"6.4.0" | "6.5.0" |
Fields that differed: changelog latestVersion
| Field | Before | After |
|---|---|---|
changelog |
"v6.3.0: New tools/netcheck.py live-net stability self-check: seven-registry probes (DOI.org/Crossref/E-utilities/arXiv/OpenAlex/S2/Unpaywall), live PubMed/Crossref verification ri | "v6.4.0: Transport layer gains TLS-fingerprint downgrade retry: TLS-class handshake timeouts retry once with a TLS1.2 context (fixes an environment-level issue where some CDN-hoste |
latestVersion |
"6.3.0" | "6.4.0" |
Fields that differed: changelog latestVersion
| Field | Before | After |
|---|---|---|
changelog |
"v6.2.0: Fixes a verified packaging defect: the MCP wheel manifest was missing the seven engine layers added by the modular refactor, so standalone uvx/pip installs could not start | "v6.3.0: New tools/netcheck.py live-net stability self-check: seven-registry probes (DOI.org/Crossref/E-utilities/arXiv/OpenAlex/S2/Unpaywall), live PubMed/Crossref verification ri |
latestVersion |
"6.2.0" | "6.3.0" |
Fields that differed: changelog latestVersion
| Field | Before | After |
|---|---|---|
changelog |
"v6.1.0: New --check-paper pre-submission audit package: one command over a full manuscript produces the verification report plus tiered audit (P0 fabrication suspects / P1 high-ri | "v6.2.0: Fixes a verified packaging defect: the MCP wheel manifest was missing the seven engine layers added by the modular refactor, so standalone uvx/pip installs could not start |
latestVersion |
"6.1.0" | "6.2.0" |
Fields that differed: changelog latestVersion
| Field | Before | After |
|---|---|---|
changelog |
"v6.0.0: The single-file engine is layered into a shell plus seven modules (text/constants, transport/state, registry lookups, semantic audit, verdict orchestration, rendering/expo | "v6.1.0: New --check-paper pre-submission audit package: one command over a full manuscript produces the verification report plus tiered audit (P0 fabrication suspects / P1 high-ri |
latestVersion |
"6.0.0" | "6.1.0" |
Fields that differed: changelog latestVersion
| Field | Before | After |
|---|---|---|
changelog |
"v5.9.0: The scholarly-usage panel gains a citation-context divergence signal: a heuristic contrastive/critical phrasing scan over sampled Semantic Scholar contexts flags disagreem | "v6.0.0: The single-file engine is layered into a shell plus seven modules (text/constants, transport/state, registry lookups, semantic audit, verdict orchestration, rendering/expo |
latestVersion |
"5.9.0" | "6.0.0" |
Fields that differed: changelog latestVersion
| Field | Before | After |
|---|---|---|
changelog |
"v4.8.0: Landing-page title cross-check for EB/OL references that land partial: a match is recorded as existence evidence in the workpaper; the release chain polling window is wide | "v5.9.0: The scholarly-usage panel gains a citation-context divergence signal: a heuristic contrastive/critical phrasing scan over sampled Semantic Scholar contexts flags disagreem |
latestVersion |
"4.8.0" | "5.9.0" |
Fields that differed: changelog latestVersion
| Field | Before | After |
|---|---|---|
changelog |
"v4.5.0: Literature discovery: give it a topic, get back a reference list where every item was registry-verified at birth (--discover searches PubMed + OpenAlex and feeds the full | "v4.8.0: Landing-page title cross-check for EB/OL references that land partial: a match is recorded as existence evidence in the workpaper; the release chain polling window is wide |
latestVersion |
"4.5.0" | "4.8.0" |
Fields that differed: changelog latestVersion
| Field | Before | After |
|---|---|---|
changelog |
"v4.4.0: Document-level reference-list completeness: --check-document now detects orphan references (listed but never cited in the body - a writing-integrity advisory, never a verd | "v4.5.0: Literature discovery: give it a topic, get back a reference list where every item was registry-verified at birth (--discover searches PubMed + OpenAlex and feeds the full |
latestVersion |
"4.4.0" | "4.5.0" |
Fields that differed: changelog latestVersion
| Field | Before | After |
|---|---|---|
changelog |
"v4.2.0: Chinese medical citation verification end-to-end: bracket-numbered GB/T 7714 entries and title-first variants now parse; inline PMID/DOI anchors are extracted and verified | "v4.4.0: Document-level reference-list completeness: --check-document now detects orphan references (listed but never cited in the body - a writing-integrity advisory, never a verd |
latestVersion |
"4.2.0" | "4.4.0" |
Fields that differed: changelog latestVersion
| Field | Before | After |
|---|---|---|
changelog |
"v3.17.0: the parameterized-release-tooling debut: unified six-point version bump, auto-enumerated test manifest, template-generated release scripts, hardlink sentinel and golden r | "v4.2.0: Chinese medical citation verification end-to-end: bracket-numbered GB/T 7714 entries and title-first variants now parse; inline PMID/DOI anchors are extracted and verified |
latestVersion |
"3.17.0" | "4.2.0" |
Fields that differed: changelog latestVersion
| Field | Before | After |
|---|---|---|
changelog |
"v3.15.0: the script-aware-guard release. The cross-language guard now works by writing system (Latin / Cyrillic / Greek / Han / Kana / Hangul) instead of CJK-only - a Russian, Gre | "v3.17.0: the parameterized-release-tooling debut: unified six-point version bump, auto-enumerated test manifest, template-generated release scripts, hardlink sentinel and golden r |
latestVersion |
"3.15.0" | "3.17.0" |
Fields that differed: changelog latestVersion
| Field | Before | After |
|---|---|---|
changelog |
"v3.13.0: the methodology-infrastructure release (borrowed from the best-run open-source skills and adapted to citation verification). references/GLOSSARY.md - an opinionated canon | "v3.15.0: the script-aware-guard release. The cross-language guard now works by writing system (Latin / Cyrillic / Greek / Han / Kana / Hangul) instead of CJK-only - a Russian, Gre |
latestVersion |
"3.13.0" | "3.15.0" |
Fields that differed: changelog latestVersion summary
| Field | Before | After |
|---|---|---|
changelog |
"v3.12.0: the submission-workbench release. --preset submission bundles the pre-submission final check into one flag (strict CI exit codes + GB-T 7714/BibTeX/CSV exports; verdict s | "v3.13.0: the methodology-infrastructure release (borrowed from the best-run open-source skills and adapted to citation verification). references/GLOSSARY.md - an opinionated canon |
latestVersion |
"3.12.0" | "3.13.0" |
summary |
"Deep research that interrogates its own sources (Verified Deep Research): calibrates scope first (3-5 sharp questions), then searches iteratively and machine-verifies every citati | "Deep research with machine-verified citations" |
Fields that differed: changelog latestVersion
| Field | Before | After |
|---|---|---|
changelog |
"v3.11.0: the bibliography-profiling-and-version-hints release. List-level fabrication fingerprint scan (pure-local, zero network): serial identifier clusters, year over-concentrat | "v3.12.0: the submission-workbench release. --preset submission bundles the pre-submission final check into one flag (strict CI exit codes + GB-T 7714/BibTeX/CSV exports; verdict s |
latestVersion |
"3.11.0" | "3.12.0" |
Fields that differed: changelog latestVersion
| Field | Before | After |
|---|---|---|
changelog |
"v3.10.0: the cross-lingual-and-clone-detection release. Cross-lingual title bridging: a Chinese original title claimed against an English registry record (the normal case for Chin | "v3.11.0: the bibliography-profiling-and-version-hints release. List-level fabrication fingerprint scan (pure-local, zero network): serial identifier clusters, year over-concentrat |
latestVersion |
"3.10.0" | "3.11.0" |
Fields that differed: changelog latestVersion summary
| Field | Before | After |
|---|---|---|
changelog |
"v3.8.0: the documentation & examples completeness release. End-to-end worked example shipped (input refs + document excerpt -> real generated report and all four export formats, u | "v3.10.0: the cross-lingual-and-clone-detection release. Cross-lingual title bridging: a Chinese original title claimed against an English registry record (the normal case for Chin |
latestVersion |
"3.8.0" | "3.10.0" |
summary |
"Deep research that interrogates its own sources (Verified Deep Research): calibrates scope first (3-5 sharp questions), then searches iteratively across sources and languages and | "Deep research that interrogates its own sources (Verified Deep Research): calibrates scope first (3-5 sharp questions), then searches iteratively and machine-verifies every citati |
Fields that differed: changelog latestVersion
| Field | Before | After |
|---|---|---|
changelog |
"v3.7.0: the context-check release. New --check-document paper.md: parses in-text citation markers ([12], [1-4], (Author, Year), doi.org links), binds each marker to the verified r | "v3.8.0: the documentation & examples completeness release. End-to-end worked example shipped (input refs + document excerpt -> real generated report and all four export formats, u |
latestVersion |
"3.7.0" | "3.8.0" |
Fields that differed: changelog latestVersion
| Field | Before | After |
|---|---|---|
changelog |
"v3.4.1: the dedup-fingerprint release. Batch deduplication now keys on identifier + normalized title fingerprint: two references sharing a URL, DOI, or PMID but carrying different | "v3.7.0: the context-check release. New --check-document paper.md: parses in-text citation markers ([12], [1-4], (Author, Year), doi.org links), binds each marker to the verified r |
latestVersion |
"3.4.1" | "3.7.0" |
Fields that differed: changelog latestVersion
| Field | Before | After |
|---|---|---|
changelog |
"v3.4.0: the metadata-completeness release. PMID-only references now get full title/journal/year cross-checks against PubMed registry data (abbreviation-aware journal matching; gen | "v3.4.1: the dedup-fingerprint release. Batch deduplication now keys on identifier + normalized title fingerprint: two references sharing a URL, DOI, or PMID but carrying different |
latestVersion |
"3.4.0" | "3.4.1" |
Fields that differed: changelog latestVersion
| Field | Before | After |
|---|---|---|
changelog |
"v3.3.0: the heterogeneous-panel release. An optional NLI third vote" | "v3.4.0: the metadata-completeness release. PMID-only references now get full title/journal/year cross-checks against PubMed registry data (abbreviation-aware journal matching; gen |
latestVersion |
"3.3.0" | "3.4.0" |
Fields that differed: changelog latestVersion
| Field | Before | After |
|---|---|---|
changelog |
"v3.2.0: the conformance release. JSON reports now carry the BLUF block as a first-class object, conformant with the published BLUF Report Specification v1.0 (five core keys + wors | "v3.3.0: the heterogeneous-panel release. An optional NLI third vote" |
latestVersion |
"3.2.0" | "3.3.0" |
Fields that differed: changelog latestVersion
| Field | Before | After |
|---|---|---|
changelog |
"v3.1.0: the closed-loop release. The L4 evidence cascade is now wired into the main flow: references the semantic layer could not resolve (not_in_source/unclear) escalate automati | "v3.2.0: the conformance release. JSON reports now carry the BLUF block as a first-class object, conformant with the published BLUF Report Specification v1.0 (five core keys + wors |
latestVersion |
"3.1.0" | "3.2.0" |
Fields that differed: changelog latestVersion
| Field | Before | After |
|---|---|---|
changelog |
"v3.0.0: the BLUF release. Every report now opens with a machine-parseable YAML block (verdict / key_numbers / blocker / next_action) plus a 5-line human TL;DR - cite-holmes define | "v3.1.0: the closed-loop release. The L4 evidence cascade is now wired into the main flow: references the semantic layer could not resolve (not_in_source/unclear) escalate automati |
latestVersion |
"3.0.0" | "3.1.0" |
Fields that differed: changelog latestVersion
| Field | Before | After |
|---|---|---|
changelog |
"v2.0.0: the MCP three-primitives release. mcp/server.py now exposes tools verify_references AND explain_verdict (plain-language verdict explanations with evidence steps), resource | "v3.0.0: the BLUF release. Every report now opens with a machine-parseable YAML block (verdict / key_numbers / blocker / next_action) plus a 5-line human TL;DR - cite-holmes define |
latestVersion |
"2.0.0" | "3.0.0" |
Fields that differed: changelog latestVersion
| Field | Before | After |
|---|---|---|
changelog |
"v1.13.0: the discoverability release. EN description rewritten around task-language roots (citation verification, verify citations, citation checker, hallucinated references, fact | "v2.0.0: the MCP three-primitives release. mcp/server.py now exposes tools verify_references AND explain_verdict (plain-language verdict explanations with evidence steps), resource |
latestVersion |
"1.13.0" | "2.0.0" |
Fields that differed: changelog latestVersion summary
| Field | Before | After |
|---|---|---|
changelog |
"v1.12.0: the trust & transparency release. Persistent disk cache (default on, 7-day TTL, local SQLite): re-running or extending a bibliography reuses stable verdicts with zero net | "v1.13.0: the discoverability release. EN description rewritten around task-language roots (citation verification, verify citations, citation checker, hallucinated references, fact |
latestVersion |
"1.12.0" | "1.13.0" |
summary |
"Deep research that interrogates its own sources (Verified Deep Research): calibrates scope first (3-5 sharp questions), plans sub-questions, searches iteratively across sources an | "Deep research that interrogates its own sources (Verified Deep Research): calibrates scope first (3-5 sharp questions), then searches iteratively across sources and languages and |
Fields that differed: changelog latestVersion
| Field | Before | After |
|---|---|---|
changelog |
"v1.11.0: the determinism release. arXiv API transient 406/403 (its anti-bot window trips even under our 3s global pacing) no longer silently skips content verification - exponenti | "v1.12.0: the trust & transparency release. Persistent disk cache (default on, 7-day TTL, local SQLite): re-running or extending a bibliography reuses stable verdicts with zero net |
latestVersion |
"1.11.0" | "1.12.0" |
Fields that differed: changelog latestVersion
| Field | Before | After |
|---|---|---|
changelog |
"v1.10.0: the speed release, targeting the top evaluation complaint (overseas-database latency on flaky networks). Parallel verification (--workers, default 4; references are indep | "v1.11.0: the determinism release. arXiv API transient 406/403 (its anti-bot window trips even under our 3s global pacing) no longer silently skips content verification - exponenti |
latestVersion |
"1.10.0" | "1.11.0" |
Fields that differed: changelog displayName latestVersion
| Field | Before | After |
|---|---|---|
changelog |
"v1.9.0: multi-source confirmation + lower friction. Semantic Scholar third source (second positive signal when DOI.org metadata is unreachable; confirmation only, never a downgrad | "v1.10.0: the speed release, targeting the top evaluation complaint (overseas-database latency on flaky networks). Parallel verification (--workers, default 4; references are indep |
displayName |
"Pkg Cite Holmes Claw" | "Cite Holmes — Deep Research × Hallucination-Free Citations" |
latestVersion |
"1.9.0" | "1.10.0" |
Fields that differed: changelog latestVersion
| Field | Before | After |
|---|---|---|
changelog |
"v1.8.1: hotfix - OpenAlex bibliographic confirmation could silently no-op (response read outside its context manager); caught by the new strict real-network acceptance gate and co | "v1.9.0: multi-source confirmation + lower friction. Semantic Scholar third source (second positive signal when DOI.org metadata is unreachable; confirmation only, never a downgrad |
latestVersion |
"1.8.1" | "1.9.0" |
Fields that differed: changelog latestVersion
| Field | Before | After |
|---|---|---|
changelog |
"v1.7.0: author-name consistency check vs DOI-registered authors (any claimed surname hit counts, full miss -> partial), capability-boundary matrix in every report (9 machine-caugh | "v1.8.1: hotfix - OpenAlex bibliographic confirmation could silently no-op (response read outside its context manager); caught by the new strict real-network acceptance gate and co |
latestVersion |
"1.7.0" | "1.8.1" |
Fields that differed: license changelog description latestVersion
| Field | Before | After |
|---|---|---|
license |
null | "MIT-0" |
changelog |
"v1.6.0: host-level circuit breaker (2 consecutive transport failures to the same host -> skip for the rest of the batch with an honest note; broken networks no longer stall the wh | "v1.7.0: author-name consistency check vs DOI-registered authors (any claimed surname hit counts, full miss -> partial), capability-boundary matrix in every report (9 machine-caugh |
description |
"Deep research that interrogates its own sources (Verified Deep Research): calibrates scope first (3-5 sharp questions), plans sub-questions, searches iteratively across sources an | null |
latestVersion |
"1.6.0" | "1.7.0" |
Fields that differed: changelog latestVersion
| Field | Before | After |
|---|---|---|
changelog |
"v1.5.0: arXiv ID metadata validation (official export.arxiv.org API, DOI-grade similarity thresholds; nonexistent IDs -> invalid, never unreachable; resolver-404 fallback), Waybac | "v1.6.0: host-level circuit breaker (2 consecutive transport failures to the same host -> skip for the rest of the batch with an honest note; broken networks no longer stall the wh |
latestVersion |
"1.5.0" | "1.6.0" |
Fields that differed: changelog description displayName latestVersion summary
| Field | Before | After |
|---|---|---|
changelog |
"v1.4.0: DOI metadata cross-validation (real-DOI wrong-paper caught as invalid), --easy one-flag mode (auto medical profile + auto BibTeX/CSV export), transient-network retry, huma | "v1.5.0: arXiv ID metadata validation (official export.arxiv.org API, DOI-grade similarity thresholds; nonexistent IDs -> invalid, never unreachable; resolver-404 fallback), Waybac |
description |
"Deep research that interrogates its own sources (Verified Deep Research): calibrates scope first (3-5 sharp questions), plans sub-questions, searches iteratively across sources an | "Deep research that interrogates its own sources (Verified Deep Research): calibrates scope first (3-5 sharp questions), plans sub-questions, searches iteratively across sources an |
displayName |
"cite-holmes" | "Pkg Cite Holmes Claw" |
latestVersion |
"1.4.0" | "1.5.0" |
summary |
"Deep research that interrogates its own sources (Verified Deep Research): calibrates scope first (3-5 sharp questions), plans sub-questions, searches iteratively across sources an | "Deep research that interrogates its own sources (Verified Deep Research): calibrates scope first (3-5 sharp questions), plans sub-questions, searches iteratively across sources an |
Fields that differed: changelog description latestVersion summary
| Field | Before | After |
|---|---|---|
changelog |
"v1.3.0: CiteScore confidence scorecard in every report (0-100 + A-D grade; verified +10 / partial +4 / unreachable 0 / unverified -2 / invalid -8); tests/ regression suite (offlin | "v1.4.0: DOI metadata cross-validation (real-DOI wrong-paper caught as invalid), --easy one-flag mode (auto medical profile + auto BibTeX/CSV export), transient-network retry, huma |
description |
"Cite Holmes — deep research that interrogates its own sources (Verified Deep Research). Calibrates scope first (asks 3-5 sharp questions), plans sub-questions, searches iterativel | "Deep research that interrogates its own sources (Verified Deep Research): calibrates scope first (3-5 sharp questions), plans sub-questions, searches iteratively across sources an |
latestVersion |
"1.3.0" | "1.4.0" |
summary |
"Cite Holmes — deep research that interrogates its own sources (Verified Deep Research). Calibrates scope first (asks 3-5 sharp questions), plans sub-questions, searches iterativel | "Deep research that interrogates its own sources (Verified Deep Research): calibrates scope first (3-5 sharp questions), plans sub-questions, searches iteratively across sources an |
Fields that differed: changelog description latestVersion summary
| Field | Before | After |
|---|---|---|
changelog |
"English SKILL.md metadata for international listing (content unchanged); reproducible planted-fakes example" | "v1.3.0: CiteScore confidence scorecard in every report (0-100 + A-D grade; verified +10 / partial +4 / unreachable 0 / unverified -2 / invalid -8); tests/ regression suite (offlin |
description |
"Cite Holmes — deep research that interrogates its own sources (Verified Deep Research). Calibrates scope first (asks 3-5 sharp questions), plans sub-questions, searches iterativel | "Cite Holmes — deep research that interrogates its own sources (Verified Deep Research). Calibrates scope first (asks 3-5 sharp questions), plans sub-questions, searches iterativel |
latestVersion |
"1.1.1" | "1.3.0" |
summary |
"Cite Holmes — deep research that interrogates its own sources (Verified Deep Research). Calibrates scope first (asks 3-5 sharp questions), plans sub-questions, searches iterativel | "Cite Holmes — deep research that interrogates its own sources (Verified Deep Research). Calibrates scope first (asks 3-5 sharp questions), plans sub-questions, searches iterativel |
Correction
If you maintain this extension and believe anything above is inaccurate, request a correction. Corrections are published, and disputed entries are marked as disputed while under review.