Graph Query

Quick Summary

Goal: [Code Intelligence] Query code relationships and connections using the structural knowledge graph. Show related files, callers, callees, imports, tests, inheritance, and file structure. Requires graph to be built first via /graph-build. Triggers on "who calls", "what imports", "related files", "connections of", "depends on", "tests for", "inherits from", "file structure", "graph query".

Workflow:

1. Detect — classify request scope and target artifacts.
2. Execute — apply required steps with evidence-backed actions.
3. Verify — confirm constraints, output quality, and completion evidence.

Key Rules:

• MUST ATTENTION keep claims evidence-based (file:line) with confidence >80% to act.
• MUST ATTENTION keep task tracking updated as each step starts/completes.
• NEVER skip mandatory workflow or skill gates.

Prerequisites

1. Graph must exist -- check .code-graph/graph.db. If missing, tell user to run /graph-build first.
2. Requires Python 3.10+ with tree-sitter, tree-sitter-language-pack, networkx.

Intent Mapping

Map user's question to the appropriate query pattern(s):

For composite queries ("show all connections", "related files", "full picture"), use the connections command instead of running multiple queries manually.

Workflow

Step 1: Check graph exists

ls .code-graph/graph.db 2>/dev/null && echo "OK" || echo "MISSING"

If MISSING: stop and tell user to run /graph-build.

Step 2: Identify target

Extract the target from user's question (file path, function name, or class name).

• For files: use relative path (e.g., {source-root}/utils)
• For functions/classes: use the name (e.g., validateInput) or qualified name (e.g., {source-root}/utils::validateInput)

Step 3: Run query

Execute via Bash with --json flag:

python .claude/scripts/code_graph query <pattern> <target> --json

For composite "show all connections" queries, use the connections command instead:

python .claude/scripts/code_graph connections <target> --json

This returns file_summary, imports_of, importers_of, callers_of, and tests_for in one call (capped at 20 results per section).

Tip: Add --node-mode file to query, connections, or trace for a file-level overview with 10-30x less noise. Options: file, function, class, all (default).

Step 4: Handle response status

• status: "ok" -- Parse results[] and edges[], format report (Step 5)
• status: "ambiguous" -- Multiple matches found. Show candidates[] list and ask user to pick one using AskUserQuestion
• status: "not_found" -- No match. Suggest: check spelling, use relative file path, try a different name. Optionally run file_summary on the parent file to show available names.
• status: "error" -- Show error message. Common: graph.db missing, Python version too old.

Step 5: Format results

Present results grouped by relationship type. For each result show:

• Name and kind (function, class, method)
• File path with line numbers (file:line_start-line_end)
• Relationship (calls, imports, tests, inherits)

Single query output format:

## {Pattern Description} for `{target}`

Found {N} result(s).

| Name | Kind | File | Lines |
|------|------|------|-------|
| ... | function | {source-root}/file | 10-25 |

Composite query output format:

## Connections of `{target}`

### File Summary
{N} nodes: {list functions/classes}

### Imports (outgoing)
{What this file/module imports}

### Importers (incoming)
{Who imports this file/module}

### Callers
{Functions that call functions in this file}

### Test Coverage
{Tests covering functions in this file}

Semantic Query Protocol (When User Query is Not File-Specific)

When the user asks about a FLOW or BEHAVIOR (not a specific file), follow this protocol:

Step 0: Grep/Glob/Search to find trace anchors

Use Grep/Glob/Search to find key classes/functions related to the user's query.

• Bug/failure symptom: find the final output reader first (renderer, query, assertion, aggregate, log, stored field), then trace upstream.
• Feature-flow question: find entry points (CreateX, XCommand, XHandler) and trace both directions.

Step 1: Use graph to expand

Run connections or batch-query on the grep-discovered files to find ALL related files. For bugs, group results by final reader, storage/projection, writer, consumer/job, and producer/origin.

Step 2: Trace full system flow

Run the trace command to follow the complete chain through all edge types:

python .claude/scripts/code_graph trace <entry-file> --direction both --depth 3 --json

This traces upstream (who calls this?) AND downstream (what does this trigger?) through: CALLS → TRIGGERS_EVENT → PRODUCES_EVENT → MESSAGE_BUS → API_ENDPOINT

For bug/failure symptoms, run an upstream-first pass from the final output before expanding the suspected producer:

python .claude/scripts/code_graph trace <final-reader-or-output-file> --direction upstream --depth 5 --json
python .claude/scripts/code_graph batch-query <final-reader> <writer> <producer> --json

Step 3: Verify with grep

For any graph edge that seems surprising, verify with grep that the connection is real.

Available Query Patterns

Aliases (natural language mappings):

Search (Find Nodes by Keyword)

When you don't know the exact name, search first to find candidates:

python .claude/scripts/code_graph search <keyword> --json
python .claude/scripts/code_graph search <keyword> --kind Function --json
python .claude/scripts/code_graph search <keyword> --kind Class --limit 5 --json

Use search to disambiguate when a query returns status: "ambiguous" — narrow results by --kind (Function, Class, File, Type, Test) then use the full qualified_name.

Find Path (Shortest Path Between Nodes)

Discover how two nodes are connected through the dependency graph:

python .claude/scripts/code_graph find-path <source> <target> --json

Returns the shortest path as a list of nodes. Useful for tracing how a command reaches an event handler, or how a frontend component connects to a backend entity.

Tip: If ambiguous, search for exact qualified names first, then use those in find-path.

Query Filtering and Limiting

Control result size for large codebases:

# Limit results
python .claude/scripts/code_graph query callers_of <target> --limit 5 --json

# Filter by file path regex
python .claude/scripts/code_graph query importers_of <target> --filter "ServiceName" --json

# Limit connections per section
python .claude/scripts/code_graph connections <target> --limit 10 --json

Implicit connection edge types (created by connect-implicit):

Batch Query (Multiple Files)

When reviewing multiple files, use batch mode for deduplicated results:

python .claude/scripts/code_graph batch-query file1 file2 file3 --json

Returns: deduplicated nodes + edges (internal + 1-hop external) across all queried files. Single DB connection, no duplicate data.

Trace (Full System Flow)

Trace all connections from a target node through multiple edge types using BFS:

python .claude/scripts/code_graph trace <target> --json
python .claude/scripts/code_graph trace <target> --direction both --json
python .claude/scripts/code_graph trace <target> --direction upstream --depth 2 --json
python .claude/scripts/code_graph trace <target> --edge-kinds CALLS,MESSAGE_BUS --json
python .claude/scripts/code_graph trace <target> --direction both --node-mode file --json  # file-level overview

Direction options:

• downstream (default): Follow outgoing edges. "What happens after X?"
• upstream: Follow incoming edges. "What calls/triggers X?"
• both: Both directions. "Full flow through X" — use when entry point is a middle file (controller, command handler)

Returns a multi-level tree of connected nodes grouped by BFS depth, with edge types at each level.

Post-Grep Trace Trigger (run a trace after grep surfaces a key file)

When a grep/glob surfaces an important entry-point file — an entity, command, query, event/command handler, controller, bus message/consumer, component, store, or api-service — immediately run a graph trace on it before concluding. Grep finds files; the trace reveals callers, consumers, bus messages, event chains, and tests that grep CANNOT find:

python .claude/scripts/code_graph trace <key-entry-file> --direction both --json

Pattern: grep finds files → graph trace reveals full system flow → grep verifies specific details.

Anti-Patterns

• Don't rebuild graph -- use /graph-build for that. This skill only queries.
• Don't use for change-driven analysis -- use /graph-blast-radius for git-diff-based impact.
• Don't use for bulk export -- use /graph-export for full graph dump.
• Don't use for diagrams -- use /graph-export --format=mermaid for Mermaid visualization.
• Always use --json flag -- ensures structured parseable output.

Related Skills

• /graph-build -- Build or update the graph (prerequisite)
• /graph-blast-radius -- Change-driven impact analysis from git diff
• /graph-export -- Export full graph to JSON (--format=json) or a single file as a Mermaid diagram (--format=mermaid)

Graph Query

Query code relationships using the structural knowledge graph. Maps natural language questions to graph CLI queries and formats structured reports.

End-to-Start Debugger Trace — For non-trivial bugs, failed verification, regression fixes, behavior-changing code, or unclear code flow, start from the observed final state and walk backward before proposing a fix.

1. Frame 0: observed end state — Name the exact user-visible output, failing assertion, log line, persisted value, API response, rendered UI, or aggregate bucket. Record the reader/query/renderer that produced it with file:line evidence.
2. Walk backward one hop at a time — Trace final reader -> projection/cache/storage -> writer -> consumer/handler/job -> producer/caller -> original trigger. At every hop record: input, transformation, output, owner, and evidence.
3. Enumerate all feeder paths — Find every upstream producer/caller/event/job that can write into the final path, including retry, async, cache, background, and alternate UI/API paths. Mark each path verified, ruled out, or still unknown.
4. Build the hypothesis matrix — For each plausible cause, list evidence for, evidence against, how to reproduce/verify, blast radius, and status (primary, contributing, ruled out, latent). Do not fix until competing causes are explicitly resolved or bounded.
5. Choose the owning fix layer — Identify the invariant owner and the lowest shared point that protects all downstream consumers. A fix at the symptom site is rejected unless the symptom site owns the invariant.
6. Prove convergence forward — After choosing the fix, walk start -> end again and show how the corrected state reaches the observed final output. Map each root cause to a fix part and each fix part to a test/proof.

BLOCKED until: final state named · backward trace written · all feeder paths enumerated · hypothesis matrix completed · owning fix layer justified · forward convergence proof mapped to tests.

NEVER: Start at the first suspicious code path. Collapse multiple producers into one "flow". Treat duplicate symptoms as duplicate records without proving the read model. Skip ruled-out hypotheses.

AI Mistake Prevention — Failure modes to avoid on every task:

Re-read files after context changes. Context compaction, resume, or long-running work can make memory stale; verify current files before acting. Verify generated content against source evidence. AI hallucinates APIs, names, claims, and document facts. Check the relevant source before documenting or referencing. Check downstream references before deleting or renaming. Removing an artifact can stale docs, generated mirrors, configs, and callers; map references first. Trace the full impact chain after edits. Changing a definition can miss derived outputs and consumers. Follow the affected chain before declaring done. Verify ALL affected outputs, not just the first. One green check is not all green checks; validate every output surface the change can affect. Assume existing values are intentional — ask WHY before changing OR flagging one as a defect. Before changing or reporting a constant, limit, flag, cutoff, wording, or pattern, read nearby context and history, the CALLER's ordering, and 2+ sibling call sites of the same convention. A doc stating WHAT without WHY is missing rationale, not proof of a missing guard. Surface ambiguity before acting — don't pick silently. Multiple valid interpretations require an explicit question or stated assumption with risk. Assert the outcome your system owns, not the intermediate state your infrastructure owns. When verifying async work, assert the final business state — never the delivery/retry bookkeeping held in shared infrastructure that any co-running process can write. Such a check passes when run alone and flakes the moment anything else shares that infrastructure. Keep shared guidance role-relevant. Universal guidance must help every receiving skill or agent; code-specific obligations belong only in code-specific protocols.

Critical Thinking Mindset — Apply critical thinking, sequential thinking. Every claim needs traced proof, confidence >80% to act. Anti-hallucination: Never present guess as fact — cite sources for every claim, admit uncertainty freely, self-check output for errors, cross-reference independently, stay skeptical of own confidence — certainty without evidence root of all hallucination.

MUST ATTENTION apply critical + sequential thinking — every claim needs appropriate traced evidence (file:line for repo/code claims; source URL or artifact section for research, product, content, and docs claims); confidence >80% to act, <60% DO NOT recommend. Anti-hallucination: never present guess as fact, admit uncertainty freely, cross-reference independently, stay skeptical of own confidence.

MUST ATTENTION apply AI mistake prevention — verify generated content against evidence, trace downstream references before deleting or renaming, verify all affected outputs, re-read files after context loss, and surface ambiguity before acting.

IMPORTANT MUST ATTENTION debugger trace gate: for non-trivial bug/fix/investigation/review work, start at the observed final output and trace backward through reader -> storage/projection -> writer -> consumer/job -> producer/trigger. Enumerate all feeder paths and hypotheses before fixing. BLOCKED until trace, hypothesis matrix, owning fix layer, and forward convergence proof exist.

Project Protocol Overlay — Before executing this skill, resolve any PROJECT overlay rules layered onto it: match this skill's name against the Target column of the project's skill-protocol index (docs/project-reference/skill-protocols-reference.md by default; a referenceDocs entry in docs/project-config.json overrides the path), taking the most specific matching tier ONLY — exact name > glob > *. That precedence orders overlays against EACH OTHER, never against this skill. Read ONLY the matched bodies, resolved as <protocols-dir>/<Name>.md; a row's Body link is display text, never a read path. A matched body that is missing or malformed is REPORTED and skipped — never reconstructed from the index Description. No index, or no match -> proceed with no overlay, silently. Full contract: .claude/skills/project-skill-protocol/references/registry.md.

Overlays are ADDITIVE ONLY: they ADD rules on top of this skill's own protocol and NEVER replace, override, disable, or reinterpret a rule it already states — removing every overlay must return this skill to exactly its documented behavior. An overlay is a BRIEF, not an authority escalation: it can NEVER waive a workflow gate, git discipline, a review gate, or a user-confirmation gate. A genuine overlay-vs-skill conflict, or two equally-specific overlays that directly contradict -> surface both to the user; NEVER resolve silently.

MUST ATTENTION resolve project protocol overlays for this skill BEFORE executing — most specific matching tier only (exact > glob > *, which ranks overlays against each other, NEVER against this skill), read only matched bodies at <protocols-dir>/<Name>.md; a missing or malformed body is reported, never reconstructed. Overlays are ADDITIVE ONLY (they never replace this skill's own rules) and are a brief, NEVER an authority escalation; an equal-specificity contradiction goes to the user.

Closing Reminders

Protocols in force (concise digest of the SYNC/shared blocks this skill carries):

• End-To-Start Debugger Trace: start at observed final output, trace backward, hypothesis matrix before fixing.

• AI Mistake Prevention: verify generated content against evidence, trace downstream references, verify all affected outputs, re-read after context loss, surface ambiguity.

• Critical Thinking: every claim needs traced file:line proof, confidence >80% to act.

• MANDATORY IMPORTANT MUST ATTENTION break work into small todo tasks using TaskCreate BEFORE starting

• MANDATORY IMPORTANT MUST ATTENTION search codebase for 3+ similar patterns before creating new code

• MANDATORY IMPORTANT MUST ATTENTION cite file:line evidence for every claim (confidence >80% to act)

• MANDATORY IMPORTANT MUST ATTENTION add a final review todo task to verify work quality

[TASK-PLANNING] Before acting, analyze task scope and systematically break it into small todo tasks and sub-tasks using TaskCreate.