While adding claude-context to the competitive comparison (README.md + generated/competitive/COMPETITIVE_ANALYSIS.md), I re-verified live GitHub stats for the tools already in the README's quick-comparison table and found the full COMPETITIVE_ANALYSIS.md (last bulk-verified 2026-03-29) is significantly out of date — not just star counts, but feature claims:
- colbymchenry/codegraph: 308 → 65,465 stars. Rewrote its core in Rust, added a native OS file-watcher for real O(changed) incremental rebuilds, expanded from ~20 to ~33-34 languages, added 9 agent-client integrations (Cursor, Codex, Gemini, Antigravity, Kiro, GitHub Copilot, etc.) — no longer "a lightweight MCP wrapper" as the doc currently characterizes it.
- code-review-graph: 4,309 → 29,471 stars. Added GraphML/Neo4j Cypher/Obsidian/SVG export and a
fail-on-risk CI gate (both previously "—" in the doc). Its own README also now admits co-change prediction currently returns 0 results on every graded commit.
- narsil-mcp: 129 → 180 stars. Added Merkle-tree incremental indexing + watch mode (previously full re-index only) and expanded CFG/DFG language coverage.
- GitNexus: 18,453 → 45,195 stars. Added an opt-in
--pdg flag (per-function CFG/PDG + taint analysis, TS/JS only) and detect_changes git-diff impact — both previously absent from the doc.
- axon: 577 → 729 stars, still no LICENSE file (license: null via GitHub API).
I only refreshed the README.md quick-comparison table (6 tools, fully re-verified against current READMEs/source as of 2026-08-08) and added a claude-context entry to COMPETITIVE_ANALYSIS.md, since a full re-score of all 87+ ranked tools was out of scope for that task. Recommend a scheduled full refresh of COMPETITIVE_ANALYSIS.md's rankings/scores given how much has changed in ~4 months.
While adding claude-context to the competitive comparison (README.md + generated/competitive/COMPETITIVE_ANALYSIS.md), I re-verified live GitHub stats for the tools already in the README's quick-comparison table and found the full COMPETITIVE_ANALYSIS.md (last bulk-verified 2026-03-29) is significantly out of date — not just star counts, but feature claims:
fail-on-riskCI gate (both previously "—" in the doc). Its own README also now admits co-change prediction currently returns 0 results on every graded commit.--pdgflag (per-function CFG/PDG + taint analysis, TS/JS only) anddetect_changesgit-diff impact — both previously absent from the doc.I only refreshed the README.md quick-comparison table (6 tools, fully re-verified against current READMEs/source as of 2026-08-08) and added a claude-context entry to COMPETITIVE_ANALYSIS.md, since a full re-score of all 87+ ranked tools was out of scope for that task. Recommend a scheduled full refresh of COMPETITIVE_ANALYSIS.md's rankings/scores given how much has changed in ~4 months.