DevLens

Analyze a repository at a local path (or GitHub URL) into a DevLens graph and persist it. Returns the graphId, commit, and compact stats — NOT the node dump. Run this once before using the query tools on a new repo.

Compare two analyzed commits and report what changed (added/removed/code-changed/score-changed nodes). For added and code-changed nodes it also computes the upstream blast radius so you can see the impact of the change set. Both commits must already be analyzed.

One-call repo architecture brief: modules, routes, stores, hooks, key flows, connections, and health (cycles + security). Replaces the 7-step orchestration of overview+subgraph+find_nodes+khop+blast_radius+get_summaries+cycles — one call instead of 7-12. The architectural backbone (modules/routes/stores/hooks/flows) is enumerated fully and never truncated; only the long-tail core-node inventory is capped to protect budget. Read-only.

Check if the analyzed graph is stale vs the current HEAD of the repo. Reports whether the working tree is dirty, whether HEAD is ahead of the last analyzed commit, and summary coverage. Also called internally by workflow tools — call this before relying on architecture_brief / security_brief / review_pr to decide whether to re-analyze. One call; read-only.

Search/filter nodes in a graph. Returns COMPACT refs (id, name, type, path, score, 1-line summary) — not source. Combine filters (AND). Use nodeIds for an exact batch fetch. Results are score-ranked and capped by `limit` (default 25).

Impact analysis: the UPSTREAM nodes that depend on the target (who calls/uses it). Answers 'if I change this node, what could break'. Each result carries its hop distance. `radius` defaults to 2; when omitted and the direct (hop-1) fanout is >=100, it returns hop-1 only and sets truncated=true — re-call with an explicit `radius` to traverse deeper, uncapped.

One-call token-budgeted context packet: keyword-seeded retrieval (no embeddings — OSS limitation; cloud adds embeddings), intent-aware traverse, and budget-capped assembly. Replaces find_nodes -> get_node -> blast_radius -> get_summaries fan-out (5-8 calls). Keyword scoring: +3 for name match, +2 for path match, +1 for business-summary match — blended 60/40 with PageRank. Falls back to central nodes on keyword miss. Pass seedNodeIds to bypass keyword search when you already have target node ids. For intent='impact' you must pass focus or seedNodeIds (else the tool errors with IMPACT_REQUIRES_FOCUS and points to review_pr). Read-only.

Graph health report: how many nodes are summarized vs structure-only, by type, plus the model and prompt version used. Use to surface 'the graph is only 60% summarized' so you can warn the user instead of silently operating on a partial graph. One call; read-only.

Dependency expansion: the DOWNSTREAM nodes the target calls/uses, out to `radius` hops. Answers 'what does this node depend on'. Same radius/cap behavior as get_blast_radius (default 2, capped at hop-1 fanout >=100 unless radius is explicit).

Full detail for ONE node: metadata, callers (who depends on it), callees (what it calls), and its technical/business/security summaries. Returns precomputed summaries (~50 tokens) instead of source — use this before opening a file. `include` selects which sections; default = everything. `edgeTypes` filters callers/callees by edge kind.

Raw source code for a node. EXPENSIVE in tokens — only call when the summary from get_node is not enough. Returns the exact line range from the analyzed commit.

List all nodes in a specific FILE or FOLDER. Pass a file path for one file's nodes, or a folder path for everything beneath it (recursive). Great for orienting in an unfamiliar area of the codebase.

High-level orientation for one repo: framework/language fingerprint, route count, and the highest-scoring (most central) nodes. Use this instead of reading package.json or scanning directories.

List nodes flagged with a security concern at or above `minSeverity` (default 'low'), ranked by severity then score. Each entry includes the security summary. Use for security review without scanning the whole codebase.

Return the cohesive cluster (feature/module) that the seed node belongs to: its sibling nodes plus the edges internal to that cluster. Use to understand the bounded context around a node without pulling the whole graph.

Batch-fetch summaries for many nodes at once. Pass the nodeIds (e.g. from a blast-radius result) and get their technical/business/security summaries in one call — far cheaper than reading each file. `include` defaults to all three.

List repositories DevLens has already analyzed. Returns each graphId, repo path, framework, and commit count. Call this first to discover what graphs exist — every other tool needs a graphId. (Also exposed as the devlens://repos resource.)

List groups of nodes that form cyclic dependencies (circular imports/calls). Useful for spotting refactor hotspots and tangled modules. Each group lists the participating nodes.

One-call onboarding skeleton: modules, entry points (routes), state (stores/hooks), key flows, a reading path, a domain glossary (from business summaries), and gotchas (cycles, high-severity security, load-bearing nodes). The graph-derived half of onboarding. The skill then reads package.json / .env.example / README from disk for setup facts and merges. Replaces the graph half of the 5-step onboard recipe. Read-only.

One-call PR review packet: diff summary, per-changed-node impact (blast radius), test coverage (incoming TESTS edges — the agent previously could not produce this at all), and security delta (new vs resolved findings). Includes a reviewer checklist. Replaces 6-10 calls with one. Requires both commits already analyzed. Read-only.

One-call prioritized security report: all findings at or above minSeverity, enriched with blast radius for high-severity issues, plus a ranked 'fixTheseFirst' list. Replaces the 3-step security-analysis recipe. Never truncates high-severity findings. Read-only.