path: root/subagents/toml_workflows.lua

-- subagents/toml_workflows.lua
--
-- Persistent TOML workflows: discovery, validation, execution, the generated
-- `/workflow:<name>` slash commands, and the model-facing `subagents.workflow`
-- tool. This is the fixed-DAG surface. Output-dependent branching and dynamic
-- fan-out stay in the Lua API (subagents/workflow.lua); TOML deliberately does
-- not grow into a programming language.
--
-- Discovery mirrors profiles: `${XDG_CONFIG_HOME:-$HOME/.config}/panto/
-- workflows/**/*.toml` first, then `<cwd>/.panto/workflows/**/*.toml`, with the
-- project layer shadowing the user layer by resolved name (the `name` field,
-- defaulting to the file stem).
--
-- Validation runs before any inference: a workflow needs a non-empty `steps`
-- array, every step needs a unique `id`, an `agent`, and a `prompt`, every
-- entry in `needs` must name a declared step, and the dependency graph must be
-- acyclic. Discovery itself never throws — an invalid file is recorded with its
-- error and registers no command, and `subagents.workflow` reports that error
-- if the model asks for the workflow by name. The same validator checks a
-- transient `steps` definition passed straight to the tool.
--
-- Execution lowers onto the Lua job primitives. Every step's prompt is its own
-- text, then the workflow input, then one labeled section per dependency in
-- `needs` order. All ready steps start at once; as each settles, any dependent
-- whose needs are now satisfied starts immediately, so unrelated branches keep
-- running. A step whose dependency did not complete is marked "skipped" and
-- never spawns, and that skip cascades transitively. Terminal steps — those no
-- other step depends on — are returned in declaration order, which keeps the
-- output stable regardless of settle order.
--
-- Edge cases: the TOML parser returns nil rather than raising for some
-- malformed documents, so a non-table parse result is treated as a parse
-- error. A workflow input may legitimately be empty (a bare `/workflow:name`
-- with no tail), which is passed through as an empty string rather than
-- rejected. A workflow whose steps are all terminal returns every step.

local workflow = require("subagents.workflow")
local paths = require("subagents.paths")

local M = {}

-- toml2lua installs its module under the name "toml", not "toml2lua".
local TOML_MODULE = "toml"

local function host()
    return require("panto").ext
end

local function load_toml()
    local ok, toml = pcall(require, TOML_MODULE)
    if not ok or type(toml) ~= "table" or type(toml.parse) ~= "function" then
        return nil, "the 'toml2lua' rock is required to read TOML workflows"
    end
    return toml
end

-- ---------------------------------------------------------------------------
-- Parsing and validation
-- ---------------------------------------------------------------------------

local function is_array(value)
    if type(value) ~= "table" then
        return false
    end
    local count = 0
    for key in pairs(value) do
        if type(key) ~= "number" then
            return false
        end
        count = count + 1
    end
    return count == #value
end

local function optional_string(value, label)
    if value == nil then
        return nil, nil
    end
    if type(value) ~= "string" or value == "" then
        return nil, label .. " must be a non-empty string when given"
    end
    return value, nil
end

-- validate(def, fallback_name) -> normalized definition | nil, err
--
-- The returned definition is a fresh table, so a caller can trust its shape:
-- { name, description, steps = { { id, agent, prompt, model, reasoning,
-- needs }, ... }, terminal = { [id] = true } }.
function M.validate(def, fallback_name)
    if type(def) ~= "table" then
        return nil, "workflow definition must be a table"
    end

    local name, err = optional_string(def.name, "`name`")
    if err then
        return nil, err
    end
    name = name or fallback_name
    if name == nil or name == "" then
        return nil, "workflow has no name"
    end

    local description
    description, err = optional_string(def.description, "`description`")
    if err then
        return nil, err
    end

    if not is_array(def.steps) or #def.steps == 0 then
        return nil, "workflow '" .. name .. "' has no `steps` array"
    end

    local steps, by_id = {}, {}
    for index, raw in ipairs(def.steps) do
        if type(raw) ~= "table" then
            return nil, string.format("workflow '%s': step %d is not a table", name, index)
        end
        local where = string.format("workflow '%s' step %d", name, index)
        if type(raw.id) ~= "string" or raw.id == "" then
            return nil, where .. ": `id` is required and must be a non-empty string"
        end
        if by_id[raw.id] then
            return nil, string.format("workflow '%s': duplicate step id '%s'", name, raw.id)
        end
        if type(raw.agent) ~= "string" or raw.agent == "" then
            return nil, string.format("workflow '%s' step '%s': `agent` is required", name, raw.id)
        end
        if type(raw.prompt) ~= "string" or raw.prompt == "" then
            return nil, string.format("workflow '%s' step '%s': `prompt` is required", name, raw.id)
        end

        local model, model_err = optional_string(raw.model, "`model`")
        if model_err then
            return nil, string.format("workflow '%s' step '%s': %s", name, raw.id, model_err)
        end
        local reasoning, reasoning_err = optional_string(raw.reasoning, "`reasoning`")
        if reasoning_err then
            return nil, string.format("workflow '%s' step '%s': %s", name, raw.id, reasoning_err)
        end

        local needs = {}
        if raw.needs ~= nil then
            if not is_array(raw.needs) then
                return nil, string.format("workflow '%s' step '%s': `needs` must be an array", name, raw.id)
            end
            for _, need in ipairs(raw.needs) do
                if type(need) ~= "string" or need == "" then
                    return nil, string.format("workflow '%s' step '%s': `needs` entries must be step ids", name, raw.id)
                end
                needs[#needs + 1] = need
            end
        end

        local step = {
            id = raw.id,
            agent = raw.agent,
            prompt = raw.prompt,
            model = model,
            reasoning = reasoning,
            needs = needs,
        }
        steps[#steps + 1] = step
        by_id[raw.id] = step
    end

    -- Dependencies must exist before the cycle walk, so an unknown name reports
    -- itself rather than looking like a broken graph.
    local terminal = {}
    for _, step in ipairs(steps) do
        terminal[step.id] = true
    end
    for _, step in ipairs(steps) do
        for _, need in ipairs(step.needs) do
            if not by_id[need] then
                return nil, string.format("workflow '%s' step '%s': unknown dependency '%s'", name, step.id, need)
            end
            terminal[need] = nil
        end
    end

    -- Iterative-free DFS with a per-node mark: "open" means the node is on the
    -- current path, so meeting it again is a cycle.
    local mark = {}
    local function visit(step, trail)
        if mark[step.id] == "done" then
            return true
        end
        if mark[step.id] == "open" then
            return false, string.format(
                "workflow '%s': dependency cycle through '%s' (%s)",
                name, step.id, table.concat(trail, " -> ") .. " -> " .. step.id)
        end
        mark[step.id] = "open"
        trail[#trail + 1] = step.id
        for _, need in ipairs(step.needs) do
            local ok, cycle_err = visit(by_id[need], trail)
            if not ok then
                return false, cycle_err
            end
        end
        trail[#trail] = nil
        mark[step.id] = "done"
        return true
    end
    for _, step in ipairs(steps) do
        local ok, cycle_err = visit(step, {})
        if not ok then
            return nil, cycle_err
        end
    end

    return {
        name = name,
        description = description,
        steps = steps,
        by_id = by_id,
        terminal = terminal,
    }
end

-- parse(text, fallback_name) -> definition | nil, err, declared_name
--
-- On failure the third value is the `name` the document declared, when it read
-- as one, so discovery can index a broken file under the name it claims rather
-- than its filename stem — otherwise a broken project file would fail to shadow
-- the user workflow of the same name and the error would go unreported.
function M.parse(text, fallback_name)
    local toml, err = load_toml()
    if not toml then
        return nil, err
    end
    local ok, parsed = pcall(toml.parse, text, { strict = true })
    if not ok then
        return nil, "invalid TOML: " .. tostring(parsed)
    end
    if type(parsed) ~= "table" then
        return nil, "invalid TOML: the document did not parse into a table"
    end
    local def, validate_err = M.validate(parsed, fallback_name)
    if def then
        return def
    end
    local declared = parsed.name
    if type(declared) ~= "string" or declared == "" then
        declared = nil
    end
    return nil, validate_err, declared
end

-- ---------------------------------------------------------------------------
-- Discovery
-- ---------------------------------------------------------------------------

-- discover() -> { list = ordered array, by_name = map, warnings = array }
--
-- Later roots (the project layer) shadow earlier ones by resolved name. An
-- unreadable or invalid file never aborts discovery: it becomes a warning, and
-- its name maps to a definition-less entry carrying the error so the tool can
-- explain the failure if the model asks for it. An invalid file shadows under
-- the name it declares (falling back to its stem only when it declares none), so
-- a broken project workflow reports its error rather than silently letting the
-- same-named user workflow run in its place.
function M.discover()
    local list, by_name, warnings = {}, {}, {}

    -- Walking is the only part that can raise (a missing luv, a hostile
    -- filesystem); a root that cannot be read contributes a warning and no
    -- workflows, so discovery as a whole keeps its "never throws" contract.
    local roots_ok, roots = pcall(paths.config_roots, "workflows")
    if not roots_ok then
        return { list = list, by_name = by_name, warnings = { tostring(roots) } }
    end

    for _, root in ipairs(roots) do
        local walk_ok, found = pcall(paths.walk, root, ".toml")
        if not walk_ok then
            warnings[#warnings + 1] = root .. ": " .. tostring(found)
            found = {}
        end
        for _, path in ipairs(found) do
            local stem = paths.stem(path)
            local text, read_err = paths.read_file(path)
            local entry
            if not text then
                entry = { name = stem, path = path, error = tostring(read_err) }
            else
                local def, err, declared = M.parse(text, stem)
                if def then
                    entry = { name = def.name, path = path, definition = def }
                else
                    entry = { name = declared or stem, path = path, error = tostring(err) }
                end
            end
            if entry.error then
                warnings[#warnings + 1] = path .. ": " .. entry.error
            end

            local existing = by_name[entry.name]
            if existing then
                for index, candidate in ipairs(list) do
                    if candidate == existing then
                        list[index] = entry
                        break
                    end
                end
            else
                list[#list + 1] = entry
            end
            by_name[entry.name] = entry
        end
    end

    return { list = list, by_name = by_name, warnings = warnings }
end

-- ---------------------------------------------------------------------------
-- Lowering onto the Lua workflow API
-- ---------------------------------------------------------------------------

local function dependency_text(result)
    if result == nil then
        return "[failed: not run]"
    end
    if result.status == "completed" then
        return workflow.output_text(result)
    end
    return "[failed: " .. tostring(result.error or result.status or "unknown") .. "]"
end

-- The exact prompt a step receives: its own text, the workflow input, then one
-- labeled section per dependency in `needs` order.
local function step_prompt(step, input, settled)
    local parts = { step.prompt, "\n\n## Workflow input\n\n", input }
    for _, need in ipairs(step.needs) do
        parts[#parts + 1] = "\n\n## Output of " .. need .. "\n\n"
        parts[#parts + 1] = dependency_text(settled[need])
    end
    return table.concat(parts)
end

M.step_prompt = step_prompt

-- lower(def) -> workflow object
function M.lower(def)
    return workflow.workflow(function(ctx, input)
        input = input or ""
        local waiting = {}
        for index, step in ipairs(def.steps) do
            waiting[index] = step
        end

        local settled = {}
        local live, live_step = {}, {}

        -- One pass may unblock another (a skip cascades to its dependents), so
        -- this repeats until nothing more can start or be skipped.
        local function advance()
            local changed = true
            while changed do
                changed = false
                local index = 1
                while index <= #waiting do
                    local step = waiting[index]
                    local ready, skip = true, false
                    for _, need in ipairs(step.needs) do
                        local result = settled[need]
                        if result == nil then
                            ready = false
                        elseif result.status ~= "completed" then
                            skip = true
                            break
                        end
                    end

                    if skip then
                        table.remove(waiting, index)
                        settled[step.id] = {
                            status = "skipped",
                            error = "skipped: a dependency did not complete",
                        }
                        changed = true
                    elseif ready then
                        table.remove(waiting, index)
                        local handle = ctx:agent({
                            agent = step.agent,
                            prompt = step_prompt(step, input, settled),
                            model = step.model,
                            reasoning = step.reasoning,
                        })
                        live[#live + 1] = handle
                        live_step[handle] = step.id
                        changed = true
                    else
                        index = index + 1
                    end
                end
            end
        end

        advance()
        while #live > 0 do
            local result, remaining = ctx:await(live, "first")
            if result == nil then
                break
            end
            local still = {}
            for _, handle in ipairs(remaining or {}) do
                still[handle] = true
            end
            for _, handle in ipairs(live) do
                if not still[handle] then
                    settled[live_step[handle]] = result
                    break
                end
            end
            live = remaining or {}
            advance()
        end

        local out = {}
        for _, step in ipairs(def.steps) do
            if def.terminal[step.id] then
                local result = settled[step.id] or { status = "skipped", error = "skipped: never started" }
                out[#out + 1] = {
                    id = step.id,
                    status = result.status,
                    output = result.output,
                    error = result.error,
                }
            end
        end
        return out
    end)
end

-- run(def, input, profiles) -> array of terminal results
function M.run(def, input, profiles)
    return workflow.execute(M.lower(def), input or "", { profiles = profiles })
end

-- ---------------------------------------------------------------------------
-- Model- and user-visible formatting
-- ---------------------------------------------------------------------------

local function format_step(result)
    return table.concat({
        "step: " .. tostring(result.id),
        "status: " .. tostring(result.status),
        "--- output ---",
        workflow.output_text(result),
    }, "\n")
end

function M.format_results(results)
    if type(results) ~= "table" or #results == 0 then
        return "The workflow produced no terminal results."
    end
    local blocks = {}
    for index, result in ipairs(results) do
        blocks[index] = format_step(result)
    end
    return table.concat(blocks, "\n\n")
end

-- ---------------------------------------------------------------------------
-- Tool and command entry points
-- ---------------------------------------------------------------------------

local registry = nil

-- The discovered set, discovered once per activation.
function M.workflows()
    if registry == nil then
        registry = M.discover()
    end
    return registry
end

local function known_names(found)
    local names = {}
    for name in pairs(found.by_name) do
        names[#names + 1] = name
    end
    if #names == 0 then
        return "(no workflows found)"
    end
    table.sort(names)
    return table.concat(names, ", ")
end

local function run_named(name, input, profiles)
    local found = M.workflows()
    local entry = found.by_name[name]
    if not entry then
        return "Error: unknown workflow '" .. tostring(name) .. "'; known: " .. known_names(found)
    end
    if not entry.definition then
        return "Error: workflow '" .. name .. "' failed to load: " .. tostring(entry.error)
    end
    local ok, results = pcall(M.run, entry.definition, input, profiles)
    if not ok then
        return "Error: " .. tostring(results)
    end
    return M.format_results(results)
end

-- The `subagents.workflow` tool: run a discovered workflow by `name`, or a
-- transient definition supplied as `steps`. Exactly one of the two.
function M.handle(input, profiles)
    if type(input) ~= "table" then
        return "Error: expected an input object"
    end
    if type(input.prompt) ~= "string" or input.prompt == "" then
        return "Error: prompt is required and must be a non-empty string"
    end

    local has_name = input.name ~= nil
    local has_steps = input.steps ~= nil
    if has_name and has_steps then
        return "Error: pass exactly one of `name` (a discovered workflow) or `steps` (a transient one), not both"
    end
    if not has_name and not has_steps then
        return "Error: pass exactly one of `name` (a discovered workflow) or `steps` (a transient one)"
    end

    if has_name then
        if type(input.name) ~= "string" or input.name == "" then
            return "Error: `name` must be a non-empty string"
        end
        return run_named(input.name, input.prompt, profiles)
    end

    local def, err = M.validate({ name = "transient", steps = input.steps }, "transient")
    if not def then
        return "Error: " .. tostring(err)
    end
    local ok, results = pcall(M.run, def, input.prompt, profiles)
    if not ok then
        return "Error: " .. tostring(results)
    end
    return M.format_results(results)
end

-- Discover the workflows and register a `/workflow:<name>` command for each
-- valid one. Invalid files register nothing; their errors stay in the
-- discovery warnings and surface through `subagents.workflow`.
function M.discover_and_register(profiles)
    registry = M.discover()
    local ext = host()
    for _, entry in ipairs(registry.list) do
        if entry.definition then
            local def = entry.definition
            ext.register_command({
                name = "workflow:" .. def.name,
                description = def.description or ("Run the " .. def.name .. " workflow."),
                handler = function(args)
                    local ok, results = pcall(M.run, def, args or "", profiles)
                    if not ok then
                        return "[workflow error: " .. tostring(results) .. "]"
                    end
                    return M.format_results(results)
                end,
            })
        end
    end
    return registry
end

return M