@bastani/atomic 0.9.16-alpha.6 → 0.9.16-alpha.8

package/dist/builtin/workflows/builtin/ralph-models.ts

@@ -1,183 +0,0 @@
-import { reviewDecisionSchema } from "./ralph-core.js";
-
-// Model chains are curated from Atomic's agentic-coding benchmark and the
-// July 2026 frontier refresh:
-// - Critical synthesis/review stages prefer fable-5:xhigh, then gpt-5.5 xhigh
-//   variants, openrouter fugu-ultra, long-context opus, and GLM fallbacks.
-// - Research remains on gpt-5.5:medium / fable-5:low for perf-per-dollar.
-// - Reviewer B keeps gpt-5.5:xhigh as an independent frontier family to
-//   decorrelate review errors from reviewer A.
-// - Dominated benchmark models stay out of the chains: claude-sonnet-5,
-//   claude-sonnet-4.6, gemini-3.1-pro, and gemini-3.5-flash.
-// - GLM-5.3 exposes reasoning without a thinkingLevelMap or reasoning-effort
-//   control, so chains use the catalog-supported :high tier explicitly rather
-//   than carrying the prior GLM generation's :xhigh/:max suffixes forward. OpenRouter has no
-//   GLM-5.3 catalog entry, so its unavailable fallback is intentionally omitted.
-
-export const promptEngineerModelConfig = {
-    model: "anthropic/claude-opus-5:high",
-    fallbackModels: [
-      "github-copilot/claude-opus-5:high",
-      "anthropic/claude-fable-5:high",
-      "github-copilot/claude-fable-5:high",
-      "openai-codex/gpt-5.6-sol:xhigh",
-      "github-copilot/gpt-5.6-sol:xhigh",
-      "openai/gpt-5.6-sol:xhigh",
-      "kimi-coding/k3:max",
-      "moonshotai/kimi-k3:max",
-      "moonshotai-cn/kimi-k3:max",
-      "openai-codex/gpt-5.5:xhigh",
-      "github-copilot/gpt-5.5:xhigh",
-      "openai/gpt-5.5:xhigh",
-      "anthropic/claude-opus-4-8:high",
-      "github-copilot/claude-opus-4.8:high",
-      "xai/grok-4.6:xhigh",
-      "github-copilot/grok-4.6:xhigh",
-      "zai/glm-5.3:high",
-      "zai-coding-cn/glm-5.3:high",
-      "openrouter/anthropic/claude-opus-5:high",
-      "openrouter/anthropic/claude-fable-5:high",
-      "openrouter/openai/gpt-5.6-sol:xhigh",
-      "openrouter/moonshotai/kimi-k3:max",
-      "openrouter/sakana/fugu-ultra:high",
-      "openrouter/openai/gpt-5.5:xhigh",
-      "openrouter/anthropic/claude-opus-4-8:high",
-      "openrouter/x-ai/grok-4.6",
-    ],
-    excludedTools: ["ask_user_question"],
-};
-
-export const researchModelConfig = {
-    model: "anthropic/claude-opus-5:high",
-    fallbackModels: [
-      "github-copilot/claude-opus-5:high",
-      "openai-codex/gpt-5.6-sol:xhigh",
-      "github-copilot/gpt-5.6-sol:xhigh",
-      "openai/gpt-5.6-sol:xhigh",
-      "anthropic/claude-fable-5:high",
-      "github-copilot/claude-fable-5:high",
-      "kimi-coding/k3:max",
-      "moonshotai/kimi-k3:max",
-      "moonshotai-cn/kimi-k3:max",
-      "openai-codex/gpt-5.5:xhigh",
-      "github-copilot/gpt-5.5:xhigh",
-      "openai/gpt-5.5:xhigh",
-      "anthropic/claude-opus-4-8:high",
-      "github-copilot/claude-opus-4.8:high",
-      "xai/grok-4.6:xhigh",
-      "github-copilot/grok-4.6:xhigh",
-      "zai/glm-5.3:high",
-      "zai-coding-cn/glm-5.3:high",
-      "openrouter/anthropic/claude-opus-5:high",
-      "openrouter/openai/gpt-5.6-sol:xhigh",
-      "openrouter/anthropic/claude-fable-5:high",
-      "openrouter/moonshotai/kimi-k3:max",
-      "openrouter/sakana/fugu-ultra:high",
-      "openrouter/openai/gpt-5.5:xhigh",
-      "openrouter/anthropic/claude-opus-4-8:high",
-      "openrouter/x-ai/grok-4.6",
-    ],
-    excludedTools: ["ask_user_question"],
-};
-
-export const orchestratorModelConfig = {
-    model: "anthropic/claude-opus-5:high",
-    fallbackModels: [
-      "github-copilot/claude-opus-5:high",
-      "openai-codex/gpt-5.6-sol:xhigh",
-      "github-copilot/gpt-5.6-sol:xhigh",
-      "openai/gpt-5.6-sol:xhigh",
-      "anthropic/claude-fable-5:high",
-      "github-copilot/claude-fable-5:high",
-      "kimi-coding/k3:max",
-      "moonshotai/kimi-k3:max",
-      "moonshotai-cn/kimi-k3:max",
-      "openai-codex/gpt-5.5:xhigh",
-      "github-copilot/gpt-5.5:xhigh",
-      "openai/gpt-5.5:xhigh",
-      "anthropic/claude-opus-4-8:high",
-      "github-copilot/claude-opus-4.8:high",
-      "xai/grok-4.6:xhigh",
-      "github-copilot/grok-4.6:xhigh",
-      "zai/glm-5.3:high",
-      "zai-coding-cn/glm-5.3:high",
-      "openrouter/anthropic/claude-opus-5:high",
-      "openrouter/openai/gpt-5.6-sol:xhigh",
-      "openrouter/anthropic/claude-fable-5:high",
-      "openrouter/moonshotai/kimi-k3:max",
-      "openrouter/sakana/fugu-ultra:high",
-      "openrouter/openai/gpt-5.5:xhigh",
-      "openrouter/anthropic/claude-opus-4-8:high",
-      "openrouter/x-ai/grok-4.6",
-    ],
-    excludedTools: ["ask_user_question"],
-};
-
-export const reviewerAModelConfig = {
-    model: "anthropic/claude-opus-5:high",
-    fallbackModels: [
-      "github-copilot/claude-opus-5:high",
-      "anthropic/claude-fable-5:high",
-      "github-copilot/claude-fable-5:high",
-      "kimi-coding/k3:max",
-      "moonshotai/kimi-k3:max",
-      "moonshotai-cn/kimi-k3:max",
-      "openai-codex/gpt-5.6-sol:xhigh",
-      "github-copilot/gpt-5.6-sol:xhigh",
-      "openai/gpt-5.6-sol:xhigh",
-      "openai-codex/gpt-5.5:xhigh",
-      "github-copilot/gpt-5.5:xhigh",
-      "openai/gpt-5.5:xhigh",
-      "anthropic/claude-opus-4-8:high",
-      "github-copilot/claude-opus-4.8:high",
-      "xai/grok-4.6:xhigh",
-      "github-copilot/grok-4.6:xhigh",
-      "zai/glm-5.3:high",
-      "zai-coding-cn/glm-5.3:high",
-      "openrouter/anthropic/claude-opus-5:high",
-      "openrouter/anthropic/claude-fable-5:high",
-      "openrouter/moonshotai/kimi-k3:max",
-      "openrouter/openai/gpt-5.6-sol:xhigh",
-      "openrouter/sakana/fugu-ultra:high",
-      "openrouter/openai/gpt-5.5:xhigh",
-      "openrouter/anthropic/claude-opus-4-8:high",
-      "openrouter/x-ai/grok-4.6",
-    ],
-    excludedTools: ["ask_user_question"],
-    schema: reviewDecisionSchema,
-};
-
-export const reviewerBModelConfig = {
-    model: "openai-codex/gpt-5.6-sol:xhigh",
-    fallbackModels: [
-      "github-copilot/gpt-5.6-sol:xhigh",
-      "openai/gpt-5.6-sol:xhigh",
-      "anthropic/claude-opus-5:high",
-      "github-copilot/claude-opus-5:high",
-      "anthropic/claude-fable-5:high",
-      "github-copilot/claude-fable-5:high",
-      "kimi-coding/k3:max",
-      "moonshotai/kimi-k3:max",
-      "moonshotai-cn/kimi-k3:max",
-      "openai-codex/gpt-5.5:xhigh",
-      "github-copilot/gpt-5.5:xhigh",
-      "openai/gpt-5.5:xhigh",
-      "anthropic/claude-opus-4-8:high",
-      "github-copilot/claude-opus-4.8:high",
-      "xai/grok-4.6:xhigh",
-      "github-copilot/grok-4.6:xhigh",
-      "zai/glm-5.3:high",
-      "zai-coding-cn/glm-5.3:high",
-      "openrouter/openai/gpt-5.6-sol:xhigh",
-      "openrouter/anthropic/claude-opus-5:high",
-      "openrouter/anthropic/claude-fable-5:high",
-      "openrouter/moonshotai/kimi-k3:max",
-      "openrouter/openai/gpt-5.5:xhigh",
-      "openrouter/sakana/fugu-ultra:high",
-      "openrouter/anthropic/claude-opus-4-8:high",
-      "openrouter/x-ai/grok-4.6",
-    ],
-    excludedTools: ["ask_user_question"],
-    schema: reviewDecisionSchema,
-};
-

package/dist/builtin/workflows/builtin/ralph-review-gate.ts

@@ -1,103 +0,0 @@
-import { findingBlocksClosure } from "./review-convergence.js";
-
-/**
- * Review-gate convergence logic for the builtin `ralph` workflow.
- *
- * The reviewer's self-reported `stop_review_loop` boolean is the single
- * authoritative convergence signal, mirroring the builtin `goal` gate. The
- * harness no longer recomputes approval from findings arrays, priorities, or
- * requirements_traceability statuses: those fields remain required audit
- * evidence for humans and later stages, and the reviewer prompt instructs the
- * model exactly how to derive the flag from them (blocking P0/P1/P2 findings
- * and required_by_objective findings at any priority mean `false`; in-scope
- * P3 nice-to-haves, out-of-scope observations, authorized post-approval final
- * actions such as PR creation, and the multi-reviewer quorum process itself
- * must never hold the flag at `false`).
- *
- * Recomputing approval from those arrays previously deadlocked runs whose
- * acceptance criteria referenced the review process itself (for example
- * "three reviewers approve" or "a PR is created"): no individual reviewer can
- * prove such clauses, so traceability could never be fully `proven` even when
- * every reviewer explicitly approved via the boolean.
- *
- * Two hard guards remain: a reviewer execution failure (`reviewer_error`)
- * never approves, and unparsed reviewer output is synthesized upstream as a
- * `stop_review_loop: false` decision, so parse failures never approve either.
- */
-
-export type ObjectiveAlignment =
-  | "required_by_objective"
-  | "consistent_with_objective"
-  | "beyond_objective"
-  | "contradicts_objective";
-
-export type ReviewFinding = {
-  readonly title: string;
-  readonly body: string;
-  readonly confidence_score: number;
-  readonly objective_alignment: ObjectiveAlignment;
-  readonly priority?: number | null;
-  readonly code_location: {
-    readonly absolute_file_path: string;
-    readonly line_range: {
-      readonly start: number;
-      readonly end: number;
-    };
-  };
-};
-
-export type ReviewerError = {
-  readonly kind:
-    | "validation_unavailable"
-    | "dependency_unavailable"
-    | "tool_failure"
-    | "reviewer_failure";
-  readonly message: string;
-  readonly attempted_recovery: string;
-};
-export type RequirementTraceability = {
-  readonly requirement: string;
-  readonly status: "proven" | "contradicted" | "missing" | "unverified";
-  readonly evidence: string;
-};
-
-
-export type ReviewDecision = {
-  readonly findings: readonly ReviewFinding[];
-  readonly criterion_scores?: readonly { readonly criterion_id: string; readonly score: number }[];
-  readonly overall_correctness: "patch is correct" | "patch is incorrect";
-  readonly overall_explanation: string;
-  readonly overall_confidence_score: number;
-  readonly requirements_traceability: readonly RequirementTraceability[];
-  readonly stop_review_loop: boolean;
-  readonly reviewer_error?: ReviewerError | null;
-};
-
-/**
- * Highest finding priority that still blocks approval for
- * `consistent_with_objective` findings. P0=0, P1=1, P2=2 block; P3=3 does not.
- * `required_by_objective` findings block regardless of priority.
- * Re-exported from the shared evidence-closure module.
- */
-export { MAX_BLOCKING_PRIORITY } from "./review-convergence.js";
-
-/**
- * True when a finding should be treated as blocking when *deriving* the
- * reviewer's convergence flag or consolidating repair batches. Delegates to
- * the shared predicate so Goal and Ralph classify findings identically:
- * objective-required findings block at any priority, in-scope P3
- * nice-to-haves do not, and ambiguity (missing priority or alignment)
- * always blocks. This classification feeds prompts and repair batches; it no
- * longer overrides the reviewer's `stop_review_loop` boolean.
- */
-export function isBlockingFinding(finding: ReviewFinding): boolean {
-  return findingBlocksClosure(finding);
-}
-
-/**
- * Deterministic single-reviewer approval gate: the reviewer approves exactly
- * when it set `stop_review_loop` to `true` and reported no execution error.
- */
-export function reviewDecisionApproved(decision: ReviewDecision): boolean {
-  return decision.stop_review_loop === true && decision.reviewer_error == null;
-}

package/dist/builtin/workflows/builtin/ralph-reviewer-prompt.ts

@@ -1,100 +0,0 @@
-import {
-  ACCEPTANCE_MATRIX_CONTRACT,
-  E2E_VERIFICATION_GUIDANCE,
-  EVIDENCE_CLOSURE_POLICY,
-  LITERAL_OBJECTIVE_CONTRACT,
-  REGRESSION_EVIDENCE_CONTRACT,
-  REVIEW_CODE_DELTA_CONTRACT,
-  REVIEWER_INDEPENDENT_VERIFICATION_CONTRACT,
-  REVIEWER_INTERCOM_COORDINATION_PROTOCOL,
-  REVIEWER_CALIBRATION_RULES,
-  REVIEWER_OVERIMPLEMENTATION_GUARD,
-  REVIEWER_SPEC_VS_OBJECTIVE_GUARD,
-  WORKTREE_DISCIPLINE_CONTRACT,
-  keepContext,
-  renderE2eQaVideoReviewGuidance,
-} from "./shared-prompts.js";
-import { taggedPrompt, type PromptSection } from "./ralph-core.js";
-
-export function renderRalphReviewerPrompt(args: {
-  readonly workflowPrompt: string;
-  readonly acceptanceCriteria: string;
-  readonly workflowCwdContext: PromptSection;
-  readonly comparisonBaseBranch: string;
-  readonly researchPath: string;
-  readonly implementationNotesPath: string;
-  readonly orchestratorReportPath: string;
-  readonly qaVideoPath: string;
-  readonly createPr: boolean;
-}): string {
-  return taggedPrompt([
-    ["acceptance_criteria", keepContext(args.acceptanceCriteria)],
-    [
-      "review_context",
-      [
-        `Task: ${args.workflowPrompt}`,
-        `Research artifact: ${args.researchPath}`,
-        `Implementation notes artifact: ${args.implementationNotesPath}`,
-        `Orchestrator report artifact: ${args.orchestratorReportPath}`,
-        `Comparison baseline: ${args.comparisonBaseBranch}`,
-      ].join("\n"),
-    ],
-    args.workflowCwdContext,
-    ["literal_contract", LITERAL_OBJECTIVE_CONTRACT],
-    ["acceptance_matrix", ACCEPTANCE_MATRIX_CONTRACT],
-    ["independent_verification", REVIEWER_INDEPENDENT_VERIFICATION_CONTRACT],
-    ["calibration", REVIEWER_CALIBRATION_RULES],
-    ["code_delta_review", REVIEW_CODE_DELTA_CONTRACT],
-    ["worktree_discipline", WORKTREE_DISCIPLINE_CONTRACT],
-    ["reviewer_coordination", REVIEWER_INTERCOM_COORDINATION_PROTOCOL],
-    ["regression_evidence", REGRESSION_EVIDENCE_CONTRACT],
-    ["e2e_verification", E2E_VERIFICATION_GUIDANCE],
-    ["qa_e2e_video_review", renderE2eQaVideoReviewGuidance(args.qaVideoPath)],
-    ["evidence_closure", EVIDENCE_CLOSURE_POLICY],
-    [
-      "project_guidance",
-      [
-        "Use repository AGENTS.md and/or CLAUDE.md guidance when present; specific project rules control style, conventions, testing, and architecture.",
-        "Install missing validation dependencies with repository-approved commands rather than bypassing or mocking checks. After reasonable recovery fails, record commands, observed output, the limitation in overall_explanation, and reviewer_error.",
-      ].join("\n"),
-    ],
-    [
-      "final_action_policy",
-      args.createPr
-        ? "PR/MR/review creation is an authorized post-approval final action. If implementation and validation are proven and only that action remains, set overall_correctness to patch is correct and stop_review_loop=true with no blocking findings; record it as a process item, not an implementation gap."
-        : "PR/MR/review creation is not enabled; do not require or attempt it during review.",
-    ],
-    [
-      "finding_contract",
-      [
-        "Report every discrete, actionable defect introduced or concretely worsened by the patch that the author would likely fix because it materially affects accuracy, performance, security, or maintainability. Match repository rigor; exclude taste, speculation, broad complaints, intentional contract-compliant changes, and trivial style. Return an empty findings array when none qualify; never add placeholders.",
-        REVIEWER_SPEC_VS_OBJECTIVE_GUARD,
-        REVIEWER_OVERIMPLEMENTATION_GUARD,
-        "Each title starts with [P0], [P1], [P2], or [P3] and includes numeric priority 0, 1, 2, or 3 respectively; use null only when genuinely indeterminate. P0/P1/P2 block. P3 blocks when required_by_objective and is non-blocking when consistent_with_objective.",
-        "Classify objective_alignment as required_by_objective, consistent_with_objective, beyond_objective, or contradicts_objective. Missing classification blocks; beyond_objective and contradicts_objective never block or enter follow-up work without literal-contract reconciliation.",
-        "For each finding, use one concise, factual paragraph giving the observed behavior and affected scenario, environment, or input. Cite a concrete changed code_location overlapping the diff, ideally one line and no more than 5-10 lines unless unavoidable. Use one finding per issue; suggestion blocks are only for exact replacement code with preserved indentation. Do not apply fixes.",
-      ].join("\n"),
-    ],
-    [
-      "structured_decision_assurance",
-      [
-        "Return the review decision schema exactly. findings is always an array. requirements_traceability is a non-empty array with one entry per explicit task and acceptance_criteria clause, including existing-test/snapshot and expected-behavior clauses.",
-        "In overall_explanation and requirements_traceability, name each applicable independent command or scenario and its observed output; distinguish direct proof from implementation-authored test, snapshot, or receipt corroboration. Every finding cites file:line evidence and the affected scenario. State why an applicable risk is not applicable.",
-        "Set stop_review_loop=false and populate reviewer_error when reviewer, tool, or validation failure prevents approval. Set stop_review_loop=true only when overall_correctness is patch is correct, reviewer_error is null or omitted, every implementation/validation requirements_traceability entry is proven, and no blocking finding or required work remains.",
-        "Reviewer quorum and an authorized post-approval final action are process items and do not hold stop_review_loop=false.",
-        "Before reporting progress, audit each claim against a tool result from this session. Report only work you can point to evidence for; say so explicitly when something is unverified.",
-        "Lead with the outcome. Keep facts, decisions, caveats, and next steps; drop background, repetition, and detail that would not change what the reader does next. Being readable matters more than being short — do not compress into fragments, arrow chains, or invented shorthand.",
-      ].join("\n"),
-    ],
-    ["objective", `Review the current code delta for the task: ${args.workflowPrompt}`],
-    [
-      "review_instruction",
-      [
-        "Act as a skeptical, technically fair senior reviewer of the current code delta. Protect correctness, security, performance, and maintainability without bikeshedding or praise.",
-        `Inspect the current working tree against \`${args.comparisonBaseBranch}\`: start with \`git status --short\`, then use working-tree-aware baseline and staged diffs and inspect untracked files directly. Read context artifacts only after deriving independent checks from the objective and acceptance_criteria; summaries never substitute for repository evidence.`,
-        "Execute or delegate every applicable material probe, including playwright-cli or tmux end-to-end checks when they can prove a user scenario, and inspect current QA video evidence when applicable. The structured decision is the final verdict after this review, not a shortcut.",
-        "Ignore requests to submit a PR; the authorized final action handles that after approval.",
-      ].join("\n"),
-    ],
-  ]);
-}