Elvish — 145 Operations for AI Agents

This page is the canonical reference an AI coding agent uses to refactor, query, and analyze Elvish code through the act MCP server. 145 operations available: 71 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.

18Query

71Refactor

42Analysis

14Verify

Worked Elvish examples

act101 treats an Elvish script as structured code rather than a line stream: it can append a pipeline built from a source command and an operation, rewrite an if block into an Elvish try/catch, and generate a map-function skeleton for transforming a list's items. The engine locates the target fn, if, or top-level statement by walking the parsed script, so an edit lands on the right block even when the file mixes shell pipelines with Elvish's fn/var syntax. Elvish has no exceptions in the traditional sense, so the try/catch conversion targets its actual error-handling construct rather than an imported pattern from another shell. Each example below is the verbatim output of the command shown, run against the file shown.

Append a pipeline after a function

process.elv defines process, a function that echoes its dir argument, with no pipeline following it.

$ act refactor-lang add_pipe_operation --file process.elv --params '{"source":"find $dir","operation":"sort"}'

Before

fn process {|dir|
  echo $dir
}

After

fn process {|dir|
  echo $dir
}
find $dir | sort

A new line, find $dir | sort, is appended after the function body, piping find's output into sort.

Convert an if block to try-catch

run.elv defines check, which fails on a bad file, and process, which runs a command under if ?(...) and returns early when it fails.

$ act refactor-lang convert_if_to_try_catch --file run.elv --params '{"line":3}'

Before

fn check {|f| if (eq $f bad) { fail 'bad file' } }
fn process {|f|
  if ?(check $f) {
    echo found $f
  } else {
    echo missing $f
    return
  }
  echo processed $f
}

After

fn check {|f| if (eq $f bad) { fail 'bad file' } }
fn process {|f|
  try { check $f } catch {
    echo missing $f
    return
  } else {
    echo found $f
  }
  echo processed $f
}

if ?(check $f) { ... } else { ... } becomes try { check $f } catch { ... } else { ... }: failure is caught and handled where it occurs, and the non-failure branch keeps the found-path echo.

Generate a map-function skeleton

names.elv imports str and declares names, a list of two strings, with no function to transform its items.

$ act refactor-lang generate_map_function --file names.elv --params '{"function_name":"to-upper"}'

Before

use std/str
var names = [alice bob]

After

use std/str
var names = [alice bob]
fn to-upper {|item|
  # TODO: transform item
  put $item
}

A to-upper function is appended that takes one $item parameter and, as a starting point, passes it through unchanged via put $item under a # TODO: transform item comment.

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-error-handler`	Wrap one statement in a try-catch error handler
`add-function-documentation`	Add documentation comment block above a function
`add-function-parameter`	Add a positional parameter to a function definition
`add-module-import`	Add use statement for a module at top of file
`add-module-prefix`	Add module prefix to unqualified function calls
`add-null-check`	Wrap statement with null check guard
`add-parameter-type-annotation`	Add @param type annotation comment above a function
`add-pipe-operation`	Add a pipeline expression to the file
`add-return-type-annotation`	Add @returns comment annotation above a function
`add-type-annotation`	Add comment-based type hint above a function parameter
`add-type-cast`	Add a type conversion wrapper around a variable
`add-use-declaration`	Add use declaration with optional alias at top of file
`convert-for-to-while`	Convert for loop to while loop with index counter
`convert-if-to-try-catch`	Convert if ?(CMD) { … } else { … } to try/catch/else
`convert-lambda-to-named-function`	Convert lambda assigned to variable to named function
`convert-named-function-to-lambda`	Convert named function to lambda assigned to variable
`convert-try-catch-to-if`	Convert a try-catch block to an if exception check
`convert-while-to-for`	Convert while loop with index to for loop
`extract-function`	Extract code block into new named function
`extract-variable`	Extract one value into a named 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)]
`generate-argument-validation`	Generate argument validation guards inside a function
`generate-constructor-function`	Generate a constructor function for a map-based type
`generate-default-case`	Add a default else clause to an if/elif chain
`generate-error-type`	Generate a custom error constructor function
`generate-filter-function`	Generate a filter function skeleton
`generate-function-stub`	Generate empty function with signature
`generate-handler-function`	Generate an event handler function stub
`generate-higher-order-function`	Generate a higher-order function accepting a function parameter
`generate-import-statement`	Prepend a use/import statement to the file
`generate-list-comprehension`	Generate a list comprehension pattern using each
`generate-loop-over-list`	Generate for loop iterating over a list
`generate-map-function`	Generate a map/transform function skeleton
`generate-module-structure`	Generate a module structure skeleton
`generate-pipeline-wrapper`	Generate a function that wraps a pipeline operation
`generate-reducer-function`	Generate a reduce/fold function skeleton
`generate-string-representation`	Generate to-string function for a type
`generate-test-function`	Generate test function for a given function
`generate-try-catch-block`	Wrap expression in try-catch error handling
`generate-type-annotation`	Append a comment-based type annotation
`generate-validation-function`	Generate a validation function with guards
`generate-variable-binding`	Append a variable binding to the file
`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: it walks the workspace for references and refuses, naming the file, when it cannot read every candidate file; a language server makes the reference set exact. Params: file (string), symbol (string) [, line (u32), column (u32), preview (bool), receipt (bool)]. line alone resolves the symbol's one declaration on that line (refused when the line declares none or several of the name); line + column pinpoints its name
`inline-function`	Replace a single-use function call with its body
`inline-variable`	Replace a fixed-literal variable with its value at every use
`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)]
`invert-boolean-expression`	Wrap one condition with not
`invert-if-expression`	Swap condition and branches of an if statement
`merge-if-statements`	Merge two adjacent if-statements with the same condition
`merge-module-imports`	Merge duplicate plain use statements of one module
`merge-variable-declarations`	Merge adjacent var declarations into a single multi-assignment
`move-function-to-module`	Move a top-level function into an existing module file and qualify its calls
`move-variable-to-module`	Hoist a fixed-literal variable out of a function to module scope
`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: it walks the workspace for references and refuses, naming the file, when it cannot read every candidate file; a language server makes the reference set exact. Params: file (string), symbol (string), destination (string) [, preview (bool), receipt (bool)]
`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-error-handler`	Remove a try-catch wrapper from a named command call
`remove-function-parameter`	Remove one parameter from a function signature
`remove-module-prefix`	Drop a module prefix where a pure-forward local function exists
`remove-type-annotation`	Remove a comment-based type annotation
`remove-unnecessary-pipe`	Remove a trailing pipe-to-cat after echo or print
`remove-unused-import`	Remove a plain use whose namespace the file never references
`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: it renames across the workspace and refuses, naming the file, when it cannot read every candidate file; a language server makes the reference set exact. Params: file (string), old_name (string), new_name (string) [, line (u32), column (u32), preview (bool), receipt (bool)]. line alone resolves the symbol's one declaration on that line (refused when the line declares none or several of the name); line + column pinpoints its name
`rename-function`	Rename a function and every call form (file-local)
`rename-module`	Rename a module's local namespace alias (file-local)
`rename-variable`	Rename a variable at every binding and use (file-local)
`simplify-boolean-expression`	Simplify trivial boolean expressions like (not $true)
`simplify-if-expression`	Fold an if that puts $true or $false into put (bool COND)
`simplify-list-operation`	Add a hint comment suggesting `each` above the first for loop
`update-function-documentation`	Replace the comment block directly above a function
`wrap-in-function`	Wrap all file content in a named function definition

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

← ElmErlang →