Impact Analysis

Use the eight-tool core profile. Treat graph results as a lower bound wherever the tool reports heuristic resolution, dynamic dispatch, or coverage gaps.

For Rust caller or impact questions, request lsp: true in analyze_codebase and make rust-analyzer available. Inspect lsp_status.state: disabled means static-only analysis; failed means analysis continued and may retain partial LSP results. Record the error and treat affected conclusions as unverified until source checks or a successful rerun support them. completed means the pass returned, not that every call resolved. Record unresolved sites and coverage gaps.

If a full-profile workflow supplies a check_security_gates receipt, inspect both gates_passed and assessment_complete. A true pass with an incomplete assessment is not security clearance: report skipped checks, unresolved symbols and warnings. This core-profile workflow does not invoke that additional tool.

Workflow

1. Call health_check. Build or refresh the graph with analyze_codebase, and keep the exact returned graph_path.
2. If an actual unified diff or base/head refs exist, call detect_changes. Do not invent a diff for a proposal that has not been implemented.
3. Use search_codebase to resolve every proposed or changed symbol to an exact qualified name. Confirm important targets with get_symbol or get_context.
4. Call get_impact for each target. Page callers through next_offset. Treat importers, users, and implementors as capped summaries; use query_graph with an explicit ORDER BY when a complete paged list matters.
5. Call query_graph with graph="missed". Inspect flagged ranges and search relevant source, including unflagged files, even when no gaps are flagged before making a negative claim about callers, dependencies, or test coverage. Clean parsing can still miss relationships.
6. Follow high-risk dependents with get_context or get_symbol. Use sibling graphs only when the user supplies or authorizes those graph paths.

Deliverable

Separate direct graph evidence from inference. Report affected symbols, callers and dependency types, communities, graph-derived processes, co-change hints, coverage gaps, and unverified risks. Preserve the tool's epistemic qualifier instead of upgrading a lower-bound result into a completeness claim.