* fix(#2590): emit Workflow scripts the Workflow tool accepts; make the backend reachable Every emitted script was rejected. Four invalid constructs, the first fatal on its own, so the Workflow backend could never dispatch a wave: 1. no `export const meta = {…}` first statement -> whole script rejected 2. resumeFromRunId("<id>") -> "resumeFromRunId is not defined". It is a Workflow TOOL INPUT parameter, not a script function. The run id still reaches the caller via summary.resumeRunId, to pass as that input. 3. budget(<n>) -> "budget is not a function". `budget` is a read-only object { total, spent(), remaining() } fed by the caller's token directive; a script cannot set it. Recorded as intent in a comment. 4. parallel(agent(…), agent(…)) -> "parallel() expects an array of functions". Now parallel([() => agent(…), …]) — passing agent() results directly also started every agent eagerly, before parallel() could bound concurrency. The single-plan stage had its own branch with the same parallel() defect; both branches are now one array-emitting path. Waves also emit phase() calls whose titles match meta.phases exactly, so progress groups correctly. Two secondary defects kept the script from ever being REACHED — which is why this shipped undetected: 5. NOTHING resolved the Agent SDK version. The fragment claimed there was "no scriptable way" to introspect it and told callers to omit the flag, so gate 5 returned agent_sdk_version_unknown on every automated run while `capability state` still reported active:true. True for bash, false for Node: the router now reads the installed @anthropic-ai/claude-agent-sdk version, walking node_modules up the tree and reading package.json directly — require.resolve throws ERR_PACKAGE_PATH_NOT_EXPORTED because the SDK's exports map does not expose ./package.json. Precedence: explicit flag > GSD_AGENT_SDK_VERSION > installed. Fail-closed is preserved; an unresolvable version still declines to inline. A too-old SDK now reports the truthful agent_sdk_version_below_floor instead of unknown. 6. The runtime fallback was `--runtime > GSD_RUNTIME > 'unknown'`, diverging from the canonical `GSD_RUNTIME > config.runtime > 'claude'`, so any invocation without --runtime reported runtime_not_claude on an ordinary Claude project. Now delegates to runtime-slash.resolveRuntime. The fragment's `${AGENT_SDK_VERSION:+--agent-sdk-version "$AGENT_SDK_VERSION"}` snippet is removed rather than repaired: it was also shell-dependent — zsh does not word-split unquoted parameter expansions, so it collapsed to a single argv element, argValue() never matched, and the run failed into the same agent_sdk_version_unknown, indistinguishable from genuinely unknown. Auto- resolution removes the need for the construct entirely. Verified with the issue's own repro: no flags now reaches the version gate; an SDK above the floor yields backend:"workflow" with a script that parses as a real ES module. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_015TCwhbMuY37DzRMCfzTABJ * fix(#2590): sync generated registry, repair sibling tests, reject duplicate wave ids Findings from the isolated review, all fixed. HIGH — gsd-core/bin/lib/capability-registry.cjs was stale, and `lint:ci` was already RED because of it. The registry embeds the fragment text INLINE, so the shipped/installed copy still taught the exact broken contract this PR fixes: the old `${AGENT_SDK_VERSION:+…}` bash line and the "OMIT the flag when unknown" guidance. Regenerated. (I had read `lint:ci` by grepping its output instead of checking its exit code, so I recorded a red chain as green — checking $? now.) HIGH — three existing tests asserted the OLD broken shape and would have failed CI; none was touched by the first commit: tests/fix-2285-claude-orchestration-wiring.test.cjs — matched resumeFromRunId("…") tests/claude-orchestration.test.cjs — .includes('budget(') tests/claude-orchestration-command-router.test.cjs — .includes('budget(') Each now asserts the corrected contract: the id/pool reaches the caller via summary, and neither construct is ever CALLED. Two sibling assertions had also gone vacuous — `.includes('resumeFromRunId')` still passed, but only because the new explanatory COMMENT contains that substring, not because anything is wired. Rewritten to assert the real property. MEDIUM — duplicate wave ids were never rejected. Plan-id uniqueness was checked within a wave, but nothing checked wave ids across waves. That was harmless before; it is not now, because each wave emits a `phase("Wave <id>")` call plus a matching meta.phases entry and the tool matches titles by exact string — two waves sharing an id would collapse into one progress group and misattribute the second wave's agents to the first. Rejected at validation, with tests either side of the boundary. MEDIUM — the fragment contradicted itself (its "Manifest construction" header still listed $AGENT_SDK_VERSION as orchestrator-built) and, more seriously, never told the orchestrator to pass summary.resumeRunId as the Workflow tool's resumeFromRunId INPUT. Since this PR moves resume from a broken in-script call to a tool-invocation input, an implementer following only the fragment would have silently regressed phase-resume to a no-op. Both fixed. MEDIUM — docs/how-to/enable-claude-orchestration-workflow-backend.md and docs/explanation/claude-orchestration-capability.md documented `resumeFromRunId("<id>")` and `budget(<tokens>)` as current correct output — teaching the bug as the feature. Updated to the real contract, including the required meta block and the thunk-array parallel() form. (The changeset is `Fixed`, so the docs gate exempts this; it is corrected because it is wrong, not because a gate demanded it.) LOW — the router's top-of-file comment still described the divergent `--runtime > GSD_RUNTIME > 'unknown'` chain as current, ninety lines above the fix; and inserting resolveInstalledAgentSdkVersion had orphaned resolveDetectionArgs' JSDoc above the wrong function. Both repaired. lint:ci now exits 0 (verified by exit code, not by reading output). Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_015TCwhbMuY37DzRMCfzTABJ * chore(#2590): backfill changeset pr number (#2681) --------- Co-authored-by: Claude Opus 5 (1M context) <noreply@anthropic.com>
513 lines
24 KiB
JavaScript
513 lines
24 KiB
JavaScript
"use strict";
|
|
/**
|
|
* Claude Orchestration Capability — Workflow-tool backend detection + emitter
|
|
*
|
|
* #1143 — adopts Claude Code's Workflow tool (the engine behind `/effort ultracode`)
|
|
* as an optional, runtime-gated parallel-execution backend for the GSD loop.
|
|
*
|
|
* This module is the pure, testable core of the capability. It owns two seams:
|
|
*
|
|
* detectWorkflowBackend({ runtimeId, hostIntegration, config, agentSdkVersion })
|
|
* → { available: boolean, backend: 'workflow'|'inline', reason: string }
|
|
* Fail-closed: every miss degrades to `inline` (today's behaviour), so the
|
|
* core loop is byte-identical unless every gate opens. This is criteria 3 + 6.
|
|
*
|
|
* emitWorkflowScript({ phaseDir, waves, runId, budgetTokens? })
|
|
* → { ok:true, script, summary } | { ok:false, reason }
|
|
* Maps GSD's wave/plan model 1:1 onto Workflow primitives:
|
|
* wave → sequential `parallel()` stage barriers,
|
|
* plan → `agent(brief, { agentType:'gsd-executor', isolation:'worktree' })`
|
|
* — UNLESS the plan's `use_worktree` is explicitly `false`, in which case
|
|
* `isolation` is omitted entirely for that plan (#2772 / #2285 finding 1:
|
|
* a submodule-touching plan must never be forced into worktree isolation
|
|
* the inline path (execute-phase.md step 2.5) would keep it out of),
|
|
* files_modified overlap → forces plans into separate sequential stages
|
|
* (the same overlap rule execute-phase already applies inline),
|
|
* resumeFromRunId → wired to the phase run id,
|
|
* budgetTokens → a shared token pool.
|
|
* The emitted script composes the SAME gsd-executor agent the inline path
|
|
* uses, with per-plan worktree isolation mirroring the inline path's own
|
|
* per-plan decision, so it produces the same artifacts/commits (criterion 2).
|
|
* It is a generated string consumed by the orchestrator; this module never
|
|
* invokes the Workflow tool itself.
|
|
*
|
|
* Design laws:
|
|
* - Gall's Law: ship a small working slice that composes existing primitives
|
|
* (gsd-executor + worktree isolation) rather than reinventing them.
|
|
* - Greenspun's Tenth Rule (cited in #1143): adopt the Workflow tool's
|
|
* barrier/pipeline/budget/resume semantics instead of hand-rolling them.
|
|
* - Postel's Law: liberal in input (missing fields → inline), conservative in
|
|
* output (workflow only when every gate opens).
|
|
* - Fail-closed: an unknown version, a missing descriptor, or a disabled
|
|
* toggle all resolve to `inline`, never to `workflow`.
|
|
*
|
|
* Zero external dependencies. Pure functions. Never throws on bad input.
|
|
*/
|
|
// ─── Constants ────────────────────────────────────────────────────────────────
|
|
/**
|
|
* The Agent SDK version that introduced the Workflow tool (#1143 prior art).
|
|
* Used as the default floor when config does not override it. A runtime reporting
|
|
* an agentSdkVersion below this cannot host the Workflow backend.
|
|
*/
|
|
const WORKFLOW_TOOL_FLOOR_VERSION = '0.3.149';
|
|
/** Closed enum for the `claude_orchestration.execution_backend` config key. */
|
|
const BACKEND_VALUES = new Set(['auto', 'workflow', 'inline']);
|
|
/** Only this runtime can host the Workflow tool (Claude Code / Agent SDK). */
|
|
const WORKFLOW_RUNTIME = 'claude';
|
|
// ─── Semver helpers ───────────────────────────────────────────────────────────
|
|
/** Official-ish strict SemVer 2.0.0 numeric triple (+ optional pre/build). */
|
|
const SEMVER_RE = /^(0|[1-9]\d*)\.(0|[1-9]\d*)\.(0|[1-9]\d*)(?:-[0-9A-Za-z.-]+)?(?:\+[0-9A-Za-z.-]+)?$/;
|
|
/** True for a syntactically valid semver string. */
|
|
function isValidSemver(s) {
|
|
return typeof s === 'string' && SEMVER_RE.test(s);
|
|
}
|
|
/**
|
|
* Compare two semver strings.
|
|
* Returns -1/0/1 in the usual sense. Garbage in either position → -1 (fail-closed:
|
|
* an unparseable version is treated as "less than" any real floor, so detection
|
|
* never accidentally enables the preview backend on an unknown SDK).
|
|
*
|
|
* Pre-release/build metadata are ignored for the comparison — only the numeric
|
|
* major.minor.patch triple participates, matching how the Workflow-tool floor is
|
|
* specified (a plain "0.3.149").
|
|
*/
|
|
function compareSemver(a, b) {
|
|
if (!isValidSemver(a) || !isValidSemver(b))
|
|
return -1;
|
|
// Split numeric triple from pre-release/build metadata.
|
|
const parseTriple = (s) => {
|
|
const core = s.split('-')[0].split('+')[0].split('.');
|
|
return [parseInt(core[0], 10), parseInt(core[1], 10), parseInt(core[2], 10)];
|
|
};
|
|
const hasPre = (s) => s.indexOf('-') !== -1;
|
|
const preIdentifiers = (s) => (s.split('-')[1] || '').split('+')[0].split('.').filter((x) => x.length > 0);
|
|
const am = parseTriple(a);
|
|
const bm = parseTriple(b);
|
|
for (let i = 0; i < 3; i++) {
|
|
if (am[i] < bm[i])
|
|
return -1;
|
|
if (am[i] > bm[i])
|
|
return 1;
|
|
}
|
|
// Numeric triple is equal. SemVer 2.0.0 §11 precedence:
|
|
// - a version WITH a pre-release tag is LOWER than the same triple WITHOUT one
|
|
// (keeps the floor fail-closed for pre-release builds of the GA floor);
|
|
// - two pre-releases of the same triple are ordered by their dot-separated
|
|
// identifiers (numeric < alphanumeric; numeric compared numerically,
|
|
// alphanumeric lexically; fewer identifiers < more).
|
|
const aPre = hasPre(a);
|
|
const bPre = hasPre(b);
|
|
if (aPre && !bPre)
|
|
return -1;
|
|
if (!aPre && bPre)
|
|
return 1;
|
|
if (aPre && bPre) {
|
|
const ai = preIdentifiers(a);
|
|
const bi = preIdentifiers(b);
|
|
const len = Math.min(ai.length, bi.length);
|
|
for (let i = 0; i < len; i++) {
|
|
const ax = ai[i];
|
|
const bx = bi[i];
|
|
const aNum = /^\d+$/.test(ax);
|
|
const bNum = /^\d+$/.test(bx);
|
|
if (aNum && bNum) {
|
|
const an = parseInt(ax, 10);
|
|
const bn = parseInt(bx, 10);
|
|
if (an < bn)
|
|
return -1;
|
|
if (an > bn)
|
|
return 1;
|
|
}
|
|
else if (aNum && !bNum) {
|
|
return -1; // numeric identifiers always lower than alphanumeric
|
|
}
|
|
else if (!aNum && bNum) {
|
|
return 1;
|
|
}
|
|
else {
|
|
if (ax < bx)
|
|
return -1;
|
|
if (ax > bx)
|
|
return 1;
|
|
}
|
|
}
|
|
if (ai.length < bi.length)
|
|
return -1;
|
|
if (ai.length > bi.length)
|
|
return 1;
|
|
}
|
|
return 0;
|
|
}
|
|
/** Inline result shorthand. */
|
|
function inline(reason, available = false) {
|
|
return { available, backend: 'inline', reason };
|
|
}
|
|
/**
|
|
* Resolve whether the Workflow-tool backend should activate.
|
|
*
|
|
* Gate ladder (all must pass for `workflow`; first miss wins, fail-closed):
|
|
* 1. capability enabled (claude_orchestration.enabled truthy)
|
|
* 2. runtime is Claude (the only runtime that exposes the Workflow tool)
|
|
* 3. execution_backend !== 'inline'
|
|
* 4. host descriptor signals nested+background dispatch (Workflow-tool capable)
|
|
* 5. agentSdkVersion is a known, valid semver
|
|
* 6. agentSdkVersion >= the configured floor (default WORKFLOW_TOOL_FLOOR_VERSION)
|
|
* 7. execution_backend === 'workflow' OR 'auto' (both reach here; 'inline' exited at 3)
|
|
*
|
|
* Never throws. Destructures defensively.
|
|
*/
|
|
function detectWorkflowBackend(input) {
|
|
if (input === null || input === undefined || typeof input !== 'object') {
|
|
return inline('capability_disabled');
|
|
}
|
|
const cfg = (input.config !== null && input.config !== undefined && typeof input.config === 'object')
|
|
? input.config
|
|
: {};
|
|
// 1. capability must be opted in (default-off — ships disabled).
|
|
if (!cfg['claude_orchestration.enabled']) {
|
|
return inline('capability_disabled');
|
|
}
|
|
// 2. only Claude can host the Workflow tool.
|
|
if (input.runtimeId !== WORKFLOW_RUNTIME) {
|
|
return inline('runtime_not_claude');
|
|
}
|
|
// 3. explicit inline opt-out short-circuits.
|
|
let backendRaw = cfg['claude_orchestration.execution_backend'];
|
|
if (typeof backendRaw !== 'string' || !BACKEND_VALUES.has(backendRaw)) {
|
|
backendRaw = 'auto';
|
|
}
|
|
if (backendRaw === 'inline') {
|
|
return inline('backend_inline');
|
|
}
|
|
// 4. the host dispatch descriptor must be the nesting-capable Claude-Code shape
|
|
// (a proxy for Workflow-tool presence). This is Claude-specific and already
|
|
// gated at step 2; `background:true` alone is true on several non-Claude hosts,
|
|
// so the proxy is only meaningful after the runtime check above. Note: this is
|
|
// NOT the canonical `shouldFlattenDispatch` rule (which keys on
|
|
// `backgroundDispatch`); the Workflow backend works precisely because a single
|
|
// tool-call orchestrates internally, sidestepping the backgroundDispatch:false
|
|
// limitation. Missing/false/foreign descriptor → fail-closed.
|
|
const hi = input.hostIntegration;
|
|
if (hi === null || hi === undefined || typeof hi !== 'object' || Array.isArray(hi)) {
|
|
return inline('workflow_tool_unavailable');
|
|
}
|
|
const dispatch = hi.dispatch;
|
|
if (typeof dispatch !== 'object' || dispatch === null || Array.isArray(dispatch)) {
|
|
return inline('workflow_tool_unavailable');
|
|
}
|
|
const nested = dispatch['nested'];
|
|
const background = dispatch['background'];
|
|
if (nested !== true || background !== true) {
|
|
return inline('workflow_tool_unavailable');
|
|
}
|
|
// 5. an unknown agentSdkVersion cannot be trusted to meet the floor.
|
|
if (!isValidSemver(input.agentSdkVersion)) {
|
|
return inline('agent_sdk_version_unknown');
|
|
}
|
|
// 6. version floor (config override > default constant).
|
|
const floorRaw = cfg['claude_orchestration.min_agent_sdk_version'];
|
|
const floor = typeof floorRaw === 'string' && isValidSemver(floorRaw) ? floorRaw : WORKFLOW_TOOL_FLOOR_VERSION;
|
|
if (compareSemver(input.agentSdkVersion, floor) < 0) {
|
|
return inline('agent_sdk_version_below_floor');
|
|
}
|
|
// 7. auto/workflow both reach the workflow backend once every gate passes.
|
|
return { available: true, backend: 'workflow', reason: 'workflow_backend_active' };
|
|
}
|
|
/**
|
|
* Partition a wave's plans into a near-minimal number of sequential stages (via
|
|
* greedy first-fit — not guaranteed optimal for arbitrary overlap graphs, but
|
|
* correct: no two plans sharing a file ever cohabit a stage) such that no two
|
|
* plans in the same stage share a modified file. Each plan goes into the earliest
|
|
* stage where it does not overlap any plan already there.
|
|
*
|
|
* A plan with an EMPTY files_modified set declares no files; it overlaps nothing
|
|
* and coalesces into stage 0 (same behavior as the inline path, which also cannot
|
|
* guard against undeclared concurrent writes — declare filesModified accurately).
|
|
*
|
|
* This is the same overlap rule execute-phase applies inline — the only difference
|
|
* is the execution vehicle (Workflow `parallel()` vs one-agent-per-message).
|
|
*/
|
|
function partitionStages(plans) {
|
|
const stages = [];
|
|
for (const plan of plans) {
|
|
const fileSet = new Set(plan.files_modified);
|
|
let placed = false;
|
|
for (const stage of stages) {
|
|
let overlap = false;
|
|
for (const f of fileSet) {
|
|
if (stage.files.has(f)) {
|
|
overlap = true;
|
|
break;
|
|
}
|
|
}
|
|
if (!overlap) {
|
|
stage.plans.push(plan);
|
|
for (const f of fileSet)
|
|
stage.files.add(f);
|
|
placed = true;
|
|
break;
|
|
}
|
|
}
|
|
if (!placed) {
|
|
stages.push({ plans: [plan], files: new Set(fileSet) });
|
|
}
|
|
}
|
|
return stages.map((s) => s.plans.map((p) => p.id));
|
|
}
|
|
/**
|
|
* Quote a free-text value for safe embedding as a JavaScript/Workflow double-quoted
|
|
* string literal. Uses JSON.stringify so every JS-relevant escape (backslash, quote,
|
|
* newline, tab, NUL, U+2028/U+2029, all control chars) is handled by the language
|
|
* itself — there is no hand-rolled escape table to drift. Returns the value already
|
|
* wrapped in its surrounding quotes.
|
|
*/
|
|
function quoteString(s) {
|
|
return JSON.stringify(s);
|
|
}
|
|
/**
|
|
* Render the `agent()` options object for a single plan — `isolation: "worktree"`
|
|
* ONLY when the plan's `use_worktree` is not explicitly `false` (#2772 / #2285
|
|
* finding 1). This is the single place that decides worktree isolation for the
|
|
* Workflow backend; it must never diverge from the inline path's per-plan gate.
|
|
*/
|
|
function agentOptions(p) {
|
|
return p.use_worktree === false
|
|
? '{ agentType: "gsd-executor" }'
|
|
: '{ agentType: "gsd-executor", isolation: "worktree" }';
|
|
}
|
|
/**
|
|
* True if `s` is a safe identifier/path token to interpolate into the generated
|
|
* script WITHOUT requiring a string-literal context — i.e. it contains no
|
|
* character that could terminate a comment line (`\n`/`\r`), break out of a
|
|
* string literal (`"` / `\`), or smuggle a NUL/control sequence. Used for
|
|
* `phaseDir`, `runId`, `wave.id`, and `plan.id`, which are identifiers/paths and
|
|
* must never legitimately contain such characters. Rejecting them at validation
|
|
* (rather than silently flattening) keeps the emitted script faithful to input.
|
|
*/
|
|
const UNSCRIPTABLE_CHAR_RE = /[\r\n"\\\x00-\x1f\x7f\u2028\u2029]/;
|
|
function isScriptableIdentifier(s) {
|
|
if (typeof s !== 'string' || s.length === 0)
|
|
return false;
|
|
return !UNSCRIPTABLE_CHAR_RE.test(s);
|
|
}
|
|
/**
|
|
* Emit a Workflow script mapping the phase's wave/plan model onto Workflow
|
|
* primitives. Pure and deterministic: identical input yields an identical string.
|
|
*
|
|
* Returns ok:false (never throws) on invalid input — empty waves, missing runId,
|
|
* a wave with no plans, etc.
|
|
*/
|
|
function emitWorkflowScript(input) {
|
|
if (input === null || input === undefined || typeof input !== 'object') {
|
|
return { ok: false, reason: 'invalid_input' };
|
|
}
|
|
const { phaseDir, waves, runId } = input;
|
|
// Identifiers/paths interpolated into the generated script must be free of any
|
|
// character that could terminate a comment, break out of a string literal, or
|
|
// smuggle control bytes — reject up front (security: #1143 review Finding 1).
|
|
if (!isScriptableIdentifier(phaseDir)) {
|
|
return { ok: false, reason: 'phaseDir must be a non-empty string without newlines/quotes/backslash/control chars' };
|
|
}
|
|
if (!isScriptableIdentifier(runId)) {
|
|
return { ok: false, reason: 'runId must be a non-empty string without newlines/quotes/backslash/control chars' };
|
|
}
|
|
if (!Array.isArray(waves) || waves.length === 0) {
|
|
return { ok: false, reason: 'waves must be a non-empty array' };
|
|
}
|
|
// Wave ids must be unique ACROSS waves, not just plan ids within one (#2590).
|
|
// Each wave emits a `phase("Wave <id>")` call plus a matching meta.phases
|
|
// entry, and the Workflow tool matches phase titles by exact string — two
|
|
// waves sharing an id would collapse into one progress group and misattribute
|
|
// every agent in the second wave to the first.
|
|
const seenWaveIds = new Set();
|
|
for (let i = 0; i < waves.length; i++) {
|
|
const w = waves[i];
|
|
if (w === null || typeof w !== 'object' || typeof w.id !== 'string') {
|
|
return { ok: false, reason: 'waves[' + i + '] must be { id, plans: non-empty[] }' };
|
|
}
|
|
if (!isScriptableIdentifier(w.id)) {
|
|
return { ok: false, reason: 'waves[' + i + '].id must not contain newlines/quotes/backslash/control chars' };
|
|
}
|
|
if (seenWaveIds.has(w.id)) {
|
|
return { ok: false, reason: 'duplicate wave id "' + w.id + '" — wave ids must be unique (phase titles must map 1:1)' };
|
|
}
|
|
seenWaveIds.add(w.id);
|
|
if (!Array.isArray(w.plans) || w.plans.length === 0) {
|
|
return { ok: false, reason: 'waves[' + i + '] must have a non-empty plans array' };
|
|
}
|
|
const seenIds = new Set();
|
|
for (let j = 0; j < w.plans.length; j++) {
|
|
const p = w.plans[j];
|
|
if (p === null || typeof p !== 'object' || typeof p.id !== 'string' || typeof p.brief !== 'string' || !Array.isArray(p.files_modified)) {
|
|
return { ok: false, reason: 'waves[' + i + '].plans[' + j + '] must be { id, brief, files_modified[] }' };
|
|
}
|
|
if (!isScriptableIdentifier(p.id)) {
|
|
return { ok: false, reason: 'waves[' + i + '].plans[' + j + '].id must not contain newlines/quotes/backslash/control chars' };
|
|
}
|
|
if (p.use_worktree !== undefined && typeof p.use_worktree !== 'boolean') {
|
|
return { ok: false, reason: 'waves[' + i + '].plans[' + j + '].use_worktree must be a boolean if present' };
|
|
}
|
|
if (seenIds.has(p.id)) {
|
|
return { ok: false, reason: 'waves[' + i + '] has duplicate plan id "' + p.id + '"' };
|
|
}
|
|
seenIds.add(p.id);
|
|
for (const f of p.files_modified) {
|
|
if (typeof f !== 'string' || f.length === 0) {
|
|
return { ok: false, reason: 'waves[' + i + '].plans[' + j + '].files_modified entries must be non-empty strings' };
|
|
}
|
|
}
|
|
}
|
|
}
|
|
const budgetTokens = (typeof input.budgetTokens === 'number' && Number.isFinite(input.budgetTokens) && input.budgetTokens > 0)
|
|
? Math.floor(input.budgetTokens)
|
|
: null;
|
|
const lines = [];
|
|
// `export const meta = {…}` MUST be the first statement in the script — the
|
|
// Workflow tool rejects the whole script otherwise (#2590). Leading comments
|
|
// are not statements, but the meta block is emitted first regardless so the
|
|
// contract holds under the strictest reading of "first statement".
|
|
//
|
|
// meta.phases must be a PURE LITERAL (no variables, calls, spreads, or
|
|
// template interpolation), and its titles are matched EXACTLY against the
|
|
// phase() calls emitted below.
|
|
lines.push('export const meta = {');
|
|
lines.push(' name: ' + quoteString('gsd-execute-' + runId) + ',');
|
|
lines.push(' description: ' + quoteString('GSD wave dispatch for ' + phaseDir) + ',');
|
|
lines.push(' phases: [');
|
|
for (const w of waves) {
|
|
lines.push(' { title: ' + quoteString('Wave ' + w.id) + ', detail: '
|
|
+ quoteString(w.plans.length + ' plan(s)') + ' },');
|
|
}
|
|
lines.push(' ],');
|
|
lines.push('}');
|
|
lines.push('');
|
|
lines.push('// GSD Workflow script — generated by the claude-orchestration capability (#1143)');
|
|
lines.push('// phase: ' + phaseDir);
|
|
lines.push('// BETA: preview-grade; on any failure the orchestrator falls back to inline dispatch.');
|
|
lines.push('// Composes the SAME gsd-executor agent as the inline path, so artifacts (SUMMARY.md)');
|
|
lines.push('// and commits are produced identically. Worktree isolation is per-plan (use_worktree)');
|
|
lines.push('// and mirrors execute-phase.md step 2.5\'s submodule gate exactly (#2772 / #2285).');
|
|
lines.push('//');
|
|
// resumeFromRunId is a Workflow TOOL INPUT parameter, not a script function —
|
|
// calling it threw "resumeFromRunId is not defined" (#2590). The run id is
|
|
// carried in summary.resumeRunId for the caller to pass as that input.
|
|
lines.push('// resume: pass ' + quoteString(runId) + ' as the Workflow tool\'s resumeFromRunId input');
|
|
lines.push('// (it is a tool parameter, NOT a script function).');
|
|
if (budgetTokens !== null) {
|
|
// `budget` is a read-only object ({ total, spent(), remaining() }) supplied
|
|
// by the caller's token directive — a script cannot SET it, and `budget(n)`
|
|
// threw "budget is not a function" (#2590). Recorded as intent only.
|
|
lines.push('// budget: ' + budgetTokens + ' output tokens intended for this run; `budget` is');
|
|
lines.push('// read-only in a Workflow script — set it via the caller\'s token directive.');
|
|
}
|
|
lines.push('');
|
|
const stagesByWave = [];
|
|
let totalPlans = 0;
|
|
for (let wi = 0; wi < waves.length; wi++) {
|
|
const wave = waves[wi];
|
|
const stages = partitionStages(wave.plans);
|
|
stagesByWave.push(stages);
|
|
totalPlans += wave.plans.length;
|
|
lines.push('// Wave ' + wave.id);
|
|
// Title must match this wave's meta.phases entry EXACTLY.
|
|
lines.push('phase(' + quoteString('Wave ' + wave.id) + ')');
|
|
for (let si = 0; si < stages.length; si++) {
|
|
const stagePlanIds = stages[si];
|
|
// Resolve back to plan objects for briefs (ids are unique within a wave — validated above).
|
|
const stagePlans = stagePlanIds.map((id) => wave.plans.find((p) => p.id === id));
|
|
if (stages.length > 1) {
|
|
lines.push('// Stage ' + si + (si > 0 ? ' (sequential — files_modified overlap)' : ''));
|
|
}
|
|
// parallel() takes an ARRAY OF THUNKS — `parallel(agent(…), agent(…))`
|
|
// threw "parallel() expects an array of functions" (#2590). Passing
|
|
// agent() results directly would also start every agent eagerly, before
|
|
// parallel() could bound concurrency.
|
|
lines.push('await parallel([');
|
|
for (const p of stagePlans) {
|
|
lines.push(' () => agent(' + quoteString(p.brief) + ', ' + agentOptions(p) + '),');
|
|
}
|
|
lines.push('])');
|
|
}
|
|
if (wi < waves.length - 1)
|
|
lines.push('');
|
|
}
|
|
lines.push('// Each agent writes SUMMARY.md on its worktree branch; commits land there');
|
|
lines.push('// and are merged by the orchestrator exactly as in inline wave dispatch.');
|
|
const script = lines.join('\n');
|
|
return {
|
|
ok: true,
|
|
script,
|
|
summary: {
|
|
waves: waves.length,
|
|
plans: totalPlans,
|
|
stagesByWave,
|
|
resumeRunId: runId,
|
|
budgetTokens,
|
|
},
|
|
};
|
|
}
|
|
/**
|
|
* #2285 — single composed decision seam for a PRE-wave dispatch-backend selector
|
|
* (e.g. the `execute:wave:pre` claude-orchestration contribution). Composes
|
|
* `detectWorkflowBackend` (gate ladder) with `emitWorkflowScript` (wave→plan
|
|
* mapping) into ONE call so the orchestrator (and its CLI wrapper,
|
|
* `claude-orchestration resolve-wave-dispatch`) never has to re-implement the
|
|
* two-step "detect, then maybe emit" sequencing.
|
|
*
|
|
* Fail-closed at every layer, matching the two composed functions:
|
|
* - `detectWorkflowBackend` resolving anything other than `'workflow'` →
|
|
* `inline` immediately; `emitWorkflowScript` is never invoked (no wasted
|
|
* work, no risk of a bad emit masking a correct inline fallback).
|
|
* - `detectWorkflowBackend` resolves `'workflow'` but `emitWorkflowScript`
|
|
* fails (`ok:false` — e.g. a malformed wave manifest) → `inline`, carrying
|
|
* the emit failure reason so the caller can surface it. Never a partial or
|
|
* broken script.
|
|
*
|
|
* This is the designated non-CLI-router, non-test caller of
|
|
* `detectWorkflowBackend` and `emitWorkflowScript` — the standalone CLI
|
|
* subcommands (`detect-backend`, `emit-workflow`) remain for inspection/
|
|
* debugging, but the orchestrator's real per-wave dispatch decision goes
|
|
* through this seam.
|
|
*
|
|
* Never throws on bad input.
|
|
*/
|
|
function resolveWaveDispatch(input) {
|
|
if (input === null || input === undefined || typeof input !== 'object') {
|
|
return { backend: 'inline', reason: 'invalid_input' };
|
|
}
|
|
const detected = detectWorkflowBackend({
|
|
runtimeId: input.runtimeId,
|
|
hostIntegration: input.hostIntegration,
|
|
config: input.config,
|
|
agentSdkVersion: input.agentSdkVersion,
|
|
});
|
|
if (detected.backend !== 'workflow') {
|
|
return { backend: 'inline', reason: detected.reason };
|
|
}
|
|
const emitted = emitWorkflowScript({
|
|
phaseDir: input.phaseDir,
|
|
waves: input.waves,
|
|
runId: input.runId,
|
|
budgetTokens: input.budgetTokens,
|
|
});
|
|
if (!emitted.ok) {
|
|
return { backend: 'inline', reason: 'emit_failed: ' + emitted.reason };
|
|
}
|
|
return {
|
|
backend: 'workflow',
|
|
reason: detected.reason,
|
|
script: emitted.script,
|
|
summary: emitted.summary,
|
|
};
|
|
}
|
|
module.exports = {
|
|
detectWorkflowBackend,
|
|
emitWorkflowScript,
|
|
resolveWaveDispatch,
|
|
compareSemver,
|
|
isValidSemver,
|
|
WORKFLOW_TOOL_FLOOR_VERSION,
|
|
BACKEND_VALUES,
|
|
WORKFLOW_RUNTIME,
|
|
};
|