Active development: 68 commits across 3 contributors
Code ActivitySource linked
19 active day(s) in chainswarm/chain-insights.
Utility
Verified BuilderCode ActivitySource linked
19 active day(s) in chainswarm/chain-insights.
Code ActivitySource linked
18 active day(s) in chainswarm/chain-insights.
View evidenceCode ActivitySource linked
18 active day(s) in chainswarm/chain-insights.
View evidenceGithub ReleaseSource verified
What's Changed * fix: make CLI command paths self-guiding by @aph5nt in https://github.com/chainswarm/chain-insights/pull/359 Full Changelog: https://github.com/chainswarm/chain-insights/compare/v0.24.1...v0.24.2
View evidenceGithub ReleaseSource verified
What's Changed * fix: clarify CLI network and tool catalogs by @aph5nt in https://github.com/chainswarm/chain-insights/pull/357 Full Changelog: https://github.com/chainswarm/chain-insights/compare/v0.24.0...v0.24.1
View evidenceGithub ReleaseSource verified
What's Changed * feat: improve first-run CLI onboarding by @aph5nt in https://github.com/chainswarm/chain-insights/pull/355 Full Changelog: https://github.com/chainswarm/chain-insights/compare/v0.23.0...v0.24.0
View evidenceGithub ReleaseSource verified
What's Changed * fix: make fresh installs use hosted Graph by default by @aph5nt in https://github.com/chainswarm/chain-insights/pull/353 Full Changelog: https://github.com/chainswarm/chain-insights/compare/v0.22.2...v0.22.3
View evidenceGithub ReleaseSource verified
What's Changed * build(deps): bump anchore/sbom-action from 0.24.0 to 0.24.1 by @dependabot[bot] in https://github.com/chainswarm/chain-insights/pull/351 * fix: align hosted MCP tests with production by @aph5nt in https://github.com/chainswarm/chain-insights/pull/352 Full Changelog: https://github.com/chainswarm/chain-insights/compare/v0.22.1...v0.22.2
View evidenceGithub ReleaseSource verified
0.8.22 - amladdressrisk label reads are deterministic: the label subset feeding the risk level is now ordered by recency before the LIMIT, so the reported level no longer varies between runs for label-heavy identities. - Risk precedence hardening: labels remain the leading signal, but a lower-severity label no longer suppresses a more severe usable ML band — the response reports the more severe band with an mllabeldivergence driver. An AML triage verdict never fails toward "looks safe". Since 0.8.19 - 0.8.20: removed the legacy --remote flag from cia mcp aml-address-risk
View evidenceGithub ReleaseSource verified
0.8.19 - The Chain Insights Graph's ML verdict now includes an explicit abstention band: risklevel = UNSCORED means the model had too little labeled graph context to stand behind a severity. Chain Insights treats it as "no stance", never low risk: amladdressrisk stops deriving a severity from an abstained ML score (falls back to label/exchange-exposure signal, adds an mlabstained driver and mlverdict: unscored, and reports level: unscored when no other signal exists), and trace/report graph payloads normalize risklevel casing (HIGH → high) so the graph app's risk borde
View evidenceGithub ReleaseSource verified
Chain Insights v0.3.9 - Adds cia update / chain-insights update backed by the public npmjs registry. - Prompts interactive cia init users when a newer npm release is available. - Publishes the current main package, including the unreleased 0.3.8 staging test-key header support and stateless ACP proxy mode. Verification before release: - npm run release:check - npm ci --ignore-scripts --audit=false --fund=false - npm run typecheck - npm run build - npm test - npm pack --silent | tar -tf - PR #63 checks passed.
View evidenceGithub ReleaseSource verified
Chain Insights 0.3.7 Docs-only release that standardizes hosted GraphRAG MCP access wording as the daily free tier. - Public access is described as 10 execution seconds per IP per UTC day for bounded graph_query reads. - Wallet users receive the daily free tier first, then continue via x402 for sustained usage. - Removes demo/trial/preview-style framing from the public docs surfaces touched in this release. Verification before release: - npm run release:check - npm test -- tests/skills-contract.test.ts tests/mcp-client.test.ts tests/cli-mcp.test.ts - GitHub PR #57 checks green: Typecheck/Te
View evidenceGithub ReleaseSource verified
Summary - Clarifies the public GraphRAG MCP demo path: 10 execution seconds per IP per UTC day for bounded single graph_query reads. - Documents that prepared wallet users receive the daily grant first, then x402 continues automatically. - Records staging UAT evidence for the free quota and free-to-paid handoff. Verification - PR #56 checks passed on main. - Main CI for c78b641 passed: Verify, Security, Scorecard, CodeQL. - Local release gate passed for 0.3.5 -> 0.3.6. - Local targeted tests passed: tests/skills-contract.test.ts, tests/mcp-client.test.ts, tests/cli-mcp.test.ts.
View evidenceGithub ReleaseSource verified
Chain Insights v0.3.5 - Fresh workspaces are now Obsidian-compatible investigation vaults by default. - Added live case vault refresh files and CLI workflow for case notes, evidence notes, entity notes, and Graph.canvas. - Repositioned docs around Obsidian-first local work while preserving cia case export / MCP case_export for public, partner, LLM Wiki, and agent handoff bundles. Verification before release: - PR #54 checks passed. - Local npm test: 43 files, 495 tests passed. - Local npm run typecheck, npm run build, and npm run release:check passed. - Clean-workspace dogfood
View evidenceGithub ReleaseSource verified
Added docs/knowledge-exports.md with install and setup instructions for Obsidian, LLM Wiki, Codex, Claude Code, ChatGPT, and portable agent usage. - Linked the knowledge export guide from README, graph tools, investigation workspace docs, and MCP proxy docs so users can go from cia case export to an Obsidian vault or LLM Wiki ingestion flow without guessing the next step.
View evidenceGithub ReleaseSource verified
Summary - Add cia case export and MCP case_export for Obsidian, LLMWiki, Codex, Claude Code, and ChatGPT-friendly case bundles. - Export manifest.chain-insights.json, graph.chain-insights.json, Graph.canvas, Markdown notes, llms.txt, LLMWIKI.md, and agent prompts. - Add private, partner, and public redaction modes for generated exports. Verification - PR #52 checks passed, including Typecheck/Test/Build/Pack, CodeQL, npm Audit, Secret Pattern Scan, OpenSSF, and Socket. - Local npm run typecheck, npm run build, npm test, npm run release:check passed. - Clean-workspace dogfood exported and
View evidenceGithub ReleaseSource verified
Chain Insights 0.2.32 - Adds shipped Memgraph Cypher examples for Chain Insights agents, including staging-tested GraphRAG MCP reads and fixed-hop traversal fallbacks. - Expands Bittensor Cypher guidance with prefix search, address-family census, and combined SS58/EVM examples under network=bittensor. - Ensures skill-local references bundles are tracked and included in the npm package.
View evidenceGithub ReleaseSource verified
Chain Insights 0.2.31 - Ship generic chain-insights-cypher and Bittensor-specific chain-insights-bittensor-cypher skills. - Document schema-aware live/archive/facts GraphRAG MCP query patterns. - Keep Bittensor SS58 and EVM-pallet addresses under network=bittensor in shipped guidance.
View evidenceGithub ReleaseSource verified
Clarified the investigation skill for Bittensor: native Substrate/SS58 5... addresses and EVM-pallet 0x... addresses are queried under the same network=bittensor; agents should not switch networks based only on address format.
View evidenceGithub ReleaseSource verified
Added usagestatus documentation and client coverage for the GraphRAG MCP public free graphquery quota. - Updated graph access guidance to call out the default 10 execution seconds per IP per UTC day, explicit query LIMIT/pagination, and paid fallback after quota exhaustion. - Let paid-mode GraphRAG MCP clients make public free calls before wallet setup, while still surfacing wallet-ready guidance when the endpoint returns payment required.
View evidenceGithub ReleaseSource verified
Chain Insights v0.2.28 - Adds chain-insights wallet import as the user-facing wallet setup command. - Replaces raw walletPrivateKey and debug-bypass guidance in missing hosted-access errors with wallet ready, wallet topup, and invited access-key set guidance. - Keeps staging as the tester activation endpoint while production is not live. Verification before release: - npx vitest run tests/mcp-client.test.ts -t "missing hosted access" - npx vitest run tests/cli.test.ts -t "wallet --help exposes" - npm test - npm run typecheck - npm run build - `npm run relea
View evidenceGithub ReleaseSource verified
Chain Insights v0.2.27 - Makes wallet activation user-facing: chain-insights wallet ready now talks about one-time payment setup instead of approval mechanics. - Adds visible wallet ready --check-only and --payment-usdc flags while keeping old approval-named aliases hidden for compatibility. - Documents staging as the only tester activation endpoint until production is live: https://staging-mcp.chain-insights.ai/mcp. Verification before release: - npm test - npm run typecheck - npm run build - npm run release:check - node bin/cli.js wallet ready --help - `npm pack --dry-ru
View evidenceGithub ReleaseSource verified
Chain Insights v0.2.26 Fixes first-run MCP client onboarding when wallet/access setup is not configured yet. - The MCP proxy now starts its local Chain Insights tool surface before paid GraphRAG MCP fetch setup succeeds. - Fresh MCP clients can complete tools/list and see help, wallet/case tools, local AML workflows including scam_topology, and public GraphRAG primitive tool names. - Graph-backed calls still require hosted access or wallet setup and return user-facing setup guidance instead of killing proxy startup. Verification: - PR #42 merged with Verify, Security, Scorecard, CodeQ
View evidenceGithub ReleaseSource verified
Chain Insights 0.2.25 Released package: [email protected] Fixed - scam_topology auto-promotion now uses a decay-calibrated confidence threshold (0.5) instead of 0.72. The old threshold was calibrated to raw confidence, while the actual candidate score is hop/value decayed; realistic victim-anchored hop-1 incident-scale edges could land around 0.5-0.6, so nothing auto-promoted. The close-hop real-value core can now promote while dust and deeper edges remain review-only. Verification - Publish workflow succeeded for commit 3f4a3be0b4546bbc886e3234826bb2efa32c63df: http
View evidenceGithub ReleaseSource verified
Release Chain Insights 0.2.24 with the published scam_topology fixes plus wallet-ready payment setup UX, capped payment approval readiness, automatic allowance-required retry, and updated user-facing payment guidance.
View evidenceWritten by Dawnscan researchers and reviewed before publishing. Commentary with sources — not the project's own claims, and not evidence of shipping.
No published notes on chain-insights yet.
Know something about chain-insights? Sign in with wallet to submit a research note. Notes are reviewed before they appear here.