Rust — 161 Operations for AI Agents
This page is the canonical reference an AI coding agent uses to refactor, query, and analyze Rust code through the act MCP server. 161 operations available: 87 refactor, 18 query, 42 analysis, 14 verification. Each operation is callable from Claude Code, Cursor, Codex, OpenCode, or any MCP-compatible agent host. Click any operation for a stable anchor link suitable for citation.
Worked Rust examples
act101 resolves Rust symbols across definitions and call sites without invoking the compiler: a rename updates every reference in one operation, an extraction lands as a `let` binding ahead of the returning expression, and the structure query maps structs and traits into the same skeleton vocabulary agents use for every other language. Each example below is the verbatim output of the command shown, run against the file shown.
Rename a function and every call site
`next_delay` reads ambiguously next to `wait_ms`; it should say what it computes.
$ act refactor rename next_delay backoff_delay --file retry.rs --line 6 --column 8
Before
pub struct RetryPolicy {
pub max_attempts: u32,
pub base_delay_ms: u64,
}
pub fn next_delay(policy: &RetryPolicy, attempt: u32) -> Option<u64> {
if attempt >= policy.max_attempts {
return None;
}
Some(policy.base_delay_ms * 2u64.pow(attempt))
}
pub fn wait_ms(policy: &RetryPolicy, attempt: u32) -> u64 {
next_delay(policy, attempt).unwrap_or(0)
}
After
pub struct RetryPolicy {
pub max_attempts: u32,
pub base_delay_ms: u64,
}
pub fn backoff_delay(policy: &RetryPolicy, attempt: u32) -> Option<u64> {
if attempt >= policy.max_attempts {
return None;
}
Some(policy.base_delay_ms * 2u64.pow(attempt))
}
pub fn wait_ms(policy: &RetryPolicy, attempt: u32) -> u64 {
backoff_delay(policy, attempt).unwrap_or(0)
}
The definition and the call inside `wait_ms` changed together — reference resolution, not text search.
Extract backoff arithmetic into a let binding
The exponential-delay expression deserves a name before jitter complicates it.
$ act refactor extract-variable --file retry.rs --start-line 10 --start-column 10 --end-line 10 --end-column 50 delay_ms
Before
pub struct RetryPolicy {
pub max_attempts: u32,
pub base_delay_ms: u64,
}
pub fn next_delay(policy: &RetryPolicy, attempt: u32) -> Option<u64> {
if attempt >= policy.max_attempts {
return None;
}
Some(policy.base_delay_ms * 2u64.pow(attempt))
}
pub fn wait_ms(policy: &RetryPolicy, attempt: u32) -> u64 {
next_delay(policy, attempt).unwrap_or(0)
}
After
pub struct RetryPolicy {
pub max_attempts: u32,
pub base_delay_ms: u64,
}
pub fn next_delay(policy: &RetryPolicy, attempt: u32) -> Option<u64> {
if attempt >= policy.max_attempts {
return None;
}
let delay_ms = policy.base_delay_ms * 2u64.pow(attempt);
Some(delay_ms)
}
pub fn wait_ms(policy: &RetryPolicy, attempt: u32) -> u64 {
next_delay(policy, attempt).unwrap_or(0)
}
The binding is introduced immediately above the `Some(...)` return and the expression text is preserved byte for byte.
Query the file's structure as a skeleton
Before touching a file, an agent reads its declaration map instead of the whole source.
$ act query skeleton skel.rs
Before
pub struct RetryPolicy {
pub max_attempts: u32,
pub base_delay_ms: u64,
}
pub trait Backoff {
fn delay_for(&self, attempt: u32) -> u64;
}
Output
{
"type": "Skeleton",
"declarations": [
{
"kind": "class",
"name": "RetryPolicy",
"range": {
"start": {
"file": "skel.rs",
"line": 1,
"column": 1,
"byte_offset": 0
},
"end": {
"file": "skel.rs",
"line": 4,
"column": 2,
"byte_offset": 81
}
},
"visibility": "public"
},
{
"kind": "interface",
"name": "Backoff",
"range": {
"start": {
"file": "skel.rs",
"line": 6,
"column": 1,
"byte_offset": 83
},
"end": {
"file": "skel.rs",
"line": 8,
"column": 2,
"byte_offset": 150
}
},
"visibility": "public"
}
],
"_meta": {
"timing_ms": 11,
"output_tokens": 107,
"naive_tokens": 37,
"tokens_saved": 0,
"savings_pct": 0.0,
"operation_class": "read",
"deterministic_lines": 0,
"errors_prevented": 0,
"session": {
"operations": 1,
"total_tokens_saved": 0,
"total_deterministic_lines": 0,
"total_errors_prevented": 0,
"agent_efficiency_ratio": 0.0,
"session_duration_ms": 11
}
}
}
The `struct` surfaces as kind `class` and the `trait` as `interface`: act101 normalizes declaration kinds across languages, so the same agent logic that navigates TypeScript or Python navigates Rust.
Query
18 query tools, the same on every supported language. Descriptions live in the shared reference: /docs/query-tools.
callers control_flow data_flow definition diagnostics effect_closure effect_summary fix_auto get_type graph import_organize interface mutations references repo_outline skeleton symbols symbols_batch
Refactor
| Operation | Description |
|---|---|
add-borrow |
Add & to convert owned value to reference |
add-clone |
Insert .clone() call to fix move conflicts |
add-dereference |
Add * to dereference pointer/reference |
add-explicit-type |
Add an explicit type annotation to a `let` binding |
add-mutable-borrow |
Change &T to &mut T |
add-numeric-separators |
Insert `_` digit-group separators into a numeric literal |
add-parameter |
Add parameter to function signature |
add-pub-to-module |
Add pub keyword to a module declaration |
change-parameter-type |
Change parameter type annotation |
change-signature |
Change function/method signature |
change-type |
Change symbol type annotation |
change-visibility |
Change symbol visibility |
combine |
Combine related declarations |
convert-if-let-to-match |
Convert if-let expression to match |
convert-match-to-if-let |
Convert match expression to if-let pattern |
convert-to-match-expr |
Convert if-else equality chain to match expression |
convert-to-match-expression |
Convert if-else chain to match expression |
delete |
Delete symbol and update references |
encapsulate |
Encapsulate field with accessors |
extract-class |
Extract methods/fields to new type |
extract-constant |
Extract value to named constant |
extract-enum-member |
Extract enum variant payload/value |
extract-function |
Extract code into a new function |
extract-interface |
Extract trait from implementation |
extract-module |
Extract declarations into a new module |
extract-type |
Extract inline type into named type |
extract-variable |
Extract expression to variable |
extract_function |
Extract a code selection into a new function — automatically infers parameters, return types, and inserts the call site. Use instead of manually cutting/pasting code. Works without LSP; LSP improves type inference. Params: file (string), new_name (string), start_line (u32), start_column (u32), end_line (u32), end_column (u32) [, preview (bool), receipt (bool)] |
extract_variable |
Extract an expression into a named variable — inserts the declaration and replaces the expression with the variable name. Works without LSP. Params: file (string), new_name (string), start_line (u32), start_column (u32), end_line (u32), end_column (u32) [, preview (bool)] |
flatten |
Flatten nested structure |
gen-default-constructor |
Generate impl Default using Default::default() for all fields |
gen-default-impl |
Generate impl Default with type-aware field defaults |
gen-doc-comment-fn |
Generate /// doc comment above a function |
gen-doc-comment-struct |
Generate /// doc comment above a struct |
gen-exhaustive-match |
Ensure match is exhaustive by adding wildcard arm |
gen-getter-one |
Generate getter for a single struct field |
gen-getters-all |
Generate getters for all struct fields |
gen-getters-setters |
Generate getters and setters for all struct fields |
gen-match-arms |
Generate match expression for all enum variants |
gen-module-doc |
Generate module-level //! doc comment |
gen-new-fn |
Generate pub fn new() constructor for struct |
gen-setter-one |
Generate setter for a single struct field |
gen-test-fn |
Generate #[test] fn test_<name>() stub |
gen-test-module |
Generate #[cfg(test)] mod tests block |
generate-accessors |
Generate getters/setters for fields |
generate-builder |
Generate builder pattern implementation |
generate-constructor |
Generate constructor from fields |
generate-equals |
Generate equality implementation |
generate-from-json |
Generate JSON deserialization code |
generate-hash |
Generate hash implementation |
generate-impl |
Generate implementation from interface/trait |
generate-mapped-type |
Generate mapped/projection type |
generate-tests |
Generate unit test skeletons |
generate-to-json |
Generate JSON serialization code |
generate-to-string |
Generate string conversion implementation |
generate-type-guard |
Generate type-check helper |
import-add |
Add missing import/use declaration |
import-alias |
Add alias to import/use declaration |
import-organize |
Organize/sort imports |
import-remove |
Remove unused import/use declaration |
inline |
Inline a variable, function, or method — replace every usage with its definition body, then remove the original. The inverse of extract. Works without LSP (single-file); LSP enables cross-file inlining. Params: file (string), symbol (string) [, line (u32), preview (bool), receipt (bool)] |
inline |
Inline symbol at all usage sites |
inline-module |
Inline module contents into parent scope |
insert_body |
Replace a function's implementation body with new code. AST-validated — rejects if the result has parse errors, so you can't accidentally break syntax. Use instead of manual text editing for function rewrites. Params: file (string), symbol (string), code (string) [, commit (bool)] |
introduce-field |
Introduce new field in type |
introduce-parameter |
Introduce parameter in signature |
introduce-variable |
Introduce variable for expression |
move |
Move symbol to another file/module |
move-definition |
Move a definition to a target file |
move_symbol |
Move a function, class, or type to a different file and automatically update all imports across the codebase. Use instead of manually cut/paste + fixing imports. Works without LSP (single-file); LSP enables cross-file import updates. Params: file (string), symbol (string), destination (string) [, preview (bool), receipt (bool)] |
promote-local-to-parameter |
Promote local variable to function parameter |
pull-up |
Pull member up hierarchy |
push-down |
Push member down hierarchy |
recipe_run |
Run a codemod recipe: declarative match → transform → optional verify across modeled grammars. Preview lists matches; apply writes with optional E7 receipts and all-or-nothing rollback. Returns a per-site report. |
remove-borrow |
Remove unnecessary & when value can be moved |
remove-parameter |
Remove parameter from function signature |
rename |
Rename a symbol and automatically update ALL references across the codebase. Safer and faster than find-and-replace — AST-aware, won't rename strings or comments. Works without LSP (single-file); LSP enables cross-file renames. Params: file (string), old_name (string), new_name (string) [, line (u32), column (u32), preview (bool), receipt (bool)] |
rename |
Rename symbol and all references |
rename-use-import |
Add as alias to a use import |
reorder-parameters |
Reorder function parameters in specified order |
split |
Split large declaration into smaller parts |
wrap-if |
Wrap code in conditional |
wrap-lock |
Wrap critical section with lock |
wrap-null-check |
Wrap code with null/None guard |
wrap-optional |
Wrap value in Optional/Option |
wrap-try-catch |
Wrap code in try/catch equivalent |
wrap-using |
Wrap resource usage in scope guard |
Analysis
42 analysis tools, the same on every supported language. Descriptions live in the shared reference: /docs/analysis-tools.
analyze_api_diff analyze_chokepoints analyze_clones analyze_clusters analyze_cohesion analyze_conformance analyze_coupling analyze_cycle_risk analyze_cycles analyze_dead_code analyze_depth analyze_entry_points analyze_export analyze_extraction analyze_fan_balance analyze_features analyze_hotspots analyze_impact analyze_inconsistencies analyze_inheritance analyze_interface_bloat analyze_interfaces analyze_layers analyze_orphan_types analyze_patterns analyze_platform_deps analyze_readiness analyze_roles analyze_seams analyze_stability analyze_surface analyze_test_gaps analyze_thickness analyze_type_completeness churn_hotspots co_change_clusters coverage_overlay ownership_map profile_overlay simulate split_module trace_overlay
Verify
14 verify tools, the same on every supported language. Descriptions live in the shared reference: /docs/verification.
bisect_regression gate generate_test_harness scan secret_surface summarize_pr taint_flow unsafe_surface verify_behavioral_equivalence verify_contract_preserved verify_diff_semantics verify_port_parity verify_side_effects verify_test_impact