All 37 tools, grouped by what they do.

Project-wide counts (files, modules, functions, variables, edges) and the most-connected modules — call this first when orienting in an unfamiliar project.

Every module in the project as [module, kind] rows, for discovering forms or modules by name.

Enumerate parseable source files under a directory before creating a project, or preview project size.

Find where a symbol is declared by name, optionally narrowed to one module.

Find every place a symbol is used — module, procedure, line, and usage kind.

Check whether and how one procedure can reach another, with the full hop chain.

Answer "what breaks if I change X?" — every procedure that transitively depends on a symbol, sorted by depth.

Overview of one module — its procedures, what it depends on, what depends on it, and its dead-code count.

A token-budgeted bundle of source and summary context around a focus symbol or module, assembled in one call instead of many small lookups.

Exact source lines from a module between a start and end line, typically after locating a spot via symbol_lookup or references.

Raw filtered node query by label, module, name, or visibility, for cases the higher-level lookup tools don’t cover.

Raw filtered edge query by relationship, source module, or target module.

Text or regex search across the whole project.

Force a fresh SCIP index run for a project instead of waiting on the batch pipeline to trigger one.

Every project this backend knows about, with id, name, and source directory — check before creating a duplicate.

Load a new project from a local directory: enumerates parseable files, parses them, and resolves the graph in one call.

Fetch one project’s record by id.

Draft or re-draft a migration plan for a project — computes units and dependency order.

Fetch a project’s current saved migration plan — units, ordering, and dependencies.

Hand-save an edited migration plan, e.g. after reordering units or breaking a dependency cycle.

Start, or resume, a project’s migration batch.

Pause a running batch after its current in-flight unit finishes.

Overall batch state plus a per-unit outcome breakdown.

Manually advance one migration unit by a single step in its snapshot → translate → shim → verify → mark → done pipeline.

One migration unit’s current step, and its verify result once it’s past verification.

A module-name → migration-status map for a project, for coloring or annotating a UI view.

List cached procedure translations, optionally filtered — e.g. to find everything still pending approval.

Low-confidence procedures still awaiting a translation answer, each with its original source, the raw mechanical output, and the flags the transpiler raised.

Read the exact original source of one procedure, typically before answering a pending translation.

Submit your own translation for a pending item. Doesn’t approve it — that’s a separate step.

Approve or reject a batch of cached translations — approval is required before a unit can pass verify.

See current Public → Private scope-narrowing recommendations for a project.

Apply the Public → Private scope narrowing surfaced by get_scope_advice.

Remove, or stub, dead-code procedures and variables. dryRun returns a diff without writing.

Bulk find/replace across a project.

Overwrite one procedure’s source directly.

Write every approved translation for a module to an output file — a bounded write that only flushes already-approved entries, never arbitrary text.