@xn-intenton-z2a/agentic-lib 7.4.8 → 7.4.9

package/src/iterate.js

@@ -1,6 +1,6 @@
 // SPDX-License-Identifier: GPL-3.0-only
 // Copyright (C) 2025-2026 Polycode Limited
-// src/iterate.js — Shared iteration loop for CLI and MCP server
+// src/iterate.js — Shared iteration loop for CLI
 //
 // Runs N cycles of maintain → transform → fix, tracking transformation cost
 // against a budget. Stops early on consecutive test passes, no-progress, or

package/src/scripts/push-to-logs.sh

@@ -11,7 +11,7 @@
 
 set -euo pipefail
 
-BRANCH="agentic-lib-logs"
+BRANCH="${LOG_BRANCH:-agentic-lib-logs}"
 MAX_RETRIES=3
 
 # Collect files that actually exist in the workspace

package/src/seeds/zero-package.json

@@ -17,7 +17,7 @@
   "author": "",
   "license": "MIT",
   "dependencies": {
-    "@xn-intenton-z2a/agentic-lib": "^7.4.8"
+    "@xn-intenton-z2a/agentic-lib": "^7.4.9"
   },
   "devDependencies": {
     "@playwright/test": "^1.58.0",

package/src/agents/agentic-lib.yml

@@ -1,66 +0,0 @@
-# Mapping for from symbolic keys to filepaths for access by agentic-lib workflows with limits and access permissions
-paths:
-  # Filepaths for elaborator workflows
-  missionFilepath:
-    path: "MISSION.md"
-  librarySourcesFilepath:
-    path: "SOURCES.md"
-    permissions: ["write"]
-    limit: 32
-  libraryDocumentsPath:
-    path: "library/"
-    permissions: ["write"]
-    limit: 32
-  featuresPath:
-    path: "features/"
-    permissions: ["write"]
-    limit: 4
-
-  # Filepaths for engineer workflows
-  contributingFilepath:
-    path: "CONTRIBUTING.md"
-  targetTestsPath:
-    path: "tests/unit/"
-    permissions: ["write"]
-  targetSourcePath:
-    path: "src/lib/"
-    permissions: ["write"]
-  targetWebPath:
-    path: "src/web/"
-    permissions: ["write"]
-  dependenciesFilepath:
-    path: "package.json"
-    permissions: ["write"]
-  documentationPath:
-    path: "docs/"
-    permissions: ["write"]
-
-  # Filepaths for maintainer workflows
-  formattingFilepath:
-    path: ".prettierrc"
-  lintingFilepath:
-    path: "eslint.config.js"
-  readmeFilepath:
-    path: "README.md"
-    permissions: ["write"]
-
-# Execution commands
-testScript: "npm ci && npm test"
-
-# How many issues should be available to be picked up?
-featureDevelopmentIssuesWipLimit: 2
-maintenanceIssuesWipLimit: 1
-
-# How many attempts should be made to work on an issue?
-attemptsPerBranch: 3
-attemptsPerIssue: 2
-
-# Repository seeding
-seeding:
-  missionFilepath: "src/seeds/zero-MISSION.md"
-  sourcePath: "src/seeds/zero-main.js"
-  testsPath: "src/seeds/zero-main.test.js"
-  dependenciesFilepath: "src/seeds/zero-package.json"
-
-intentionBot:
-  intentionFilepath: "intentïon.md"

package/src/copilot/context.js

@@ -1,457 +0,0 @@
-// SPDX-License-Identifier: GPL-3.0-only
-// Copyright (C) 2025-2026 Polycode Limited
-// src/copilot/context.js — Context gathering and user prompt assembly
-//
-// Builds user prompts for each agent type from available local and GitHub context.
-// Works with or without GitHub data — local-only context is always sufficient.
-
-import { resolve } from "path";
-import { execSync } from "child_process";
-import { scanDirectory, readOptionalFile, extractFeatureSummary, formatPathsSection, summariseIssue, filterIssues } from "./session.js";
-import { readCumulativeCost } from "./telemetry.js";
-import { defaultLogger } from "./logger.js";
-
-/**
- * Per-agent refinements that control context precision and prompt quality.
- * These restore the tight scoping that existed in the old per-task handlers.
- */
-const AGENT_REFINEMENTS = {
-  "agent-issue-resolution": {
-    sortFeatures: "incomplete-first",
-    includeWebFiles: true,
-    highlightTargetIssue: true,
-    trackPromptBudget: true,
-  },
-  "agent-apply-fix": {
-    emphasizeTestOutput: true,
-  },
-  "agent-maintain-features": {
-    sortFeatures: "incomplete-first",
-    injectLimit: "features",
-  },
-  "agent-maintain-library": {
-    checkSourcesForUrls: true,
-    injectLimit: "library",
-  },
-};
-
-/**
- * Context requirements per agent. Defines what context each agent needs.
- * All fields are optional — the builder includes whatever is available.
- */
-const AGENT_CONTEXT = {
-  "agent-iterate": { mission: true, source: true, tests: true, features: true },
-  "agent-discovery": { source: true, tests: true },
-  "agent-issue-resolution": { mission: true, source: true, tests: true, features: true, issues: true },
-  "agent-apply-fix": { source: true, tests: true },
-  "agent-maintain-features": { mission: true, features: true, issues: true },
-  "agent-maintain-library": { library: true, librarySources: true },
-  "agent-ready-issue": { mission: true, features: true, issues: true },
-  "agent-review-issue": { source: true, tests: true, issues: true },
-  "agent-discussion-bot": { mission: true, features: true },
-  "agent-supervisor": { mission: true, features: true, issues: true },
-  "agent-director": { mission: true, features: true, issues: true, source: true, tests: true },
-};
-
-/**
- * Gather local context from the workspace filesystem.
- *
- * @param {string} workspacePath - Path to the workspace
- * @param {Object} config - Parsed agentic config (from config.js)
- * @param {Object} [options]
- * @param {Object} [options.logger]
- * @returns {Object} Context object with all available local data
- */
-export function gatherLocalContext(workspacePath, config, { logger = defaultLogger } = {}) {
-  const wsPath = resolve(workspacePath);
-  const paths = config.paths || {};
-  const tuning = config.tuning || {};
-
-  const context = {};
-
-  // Mission
-  const missionPath = paths.mission?.path || "MISSION.md";
-  context.mission = readOptionalFile(resolve(wsPath, missionPath));
-
-  // Source files
-  const sourcePath = paths.source?.path || "src/lib/";
-  const sourceDir = resolve(wsPath, sourcePath);
-  context.sourceFiles = scanDirectory(sourceDir, [".js", ".ts", ".mjs", ".cjs"], {
-    fileLimit: tuning.sourceScan || 10,
-    contentLimit: tuning.sourceContent || 5000,
-    sortByMtime: true,
-    clean: true,
-    outline: true,
-  }, logger);
-
-  // Test files
-  const testsPath = paths.tests?.path || "tests/";
-  const testsDir = resolve(wsPath, testsPath);
-  context.testFiles = scanDirectory(testsDir, [".js", ".ts", ".test.js", ".test.ts", ".spec.js"], {
-    fileLimit: tuning.sourceScan || 10,
-    contentLimit: tuning.testContent || 3000,
-    sortByMtime: true,
-    clean: true,
-  }, logger);
-
-  // Features
-  const featuresPath = paths.features?.path || "features/";
-  const featuresDir = resolve(wsPath, featuresPath);
-  const featureFiles = scanDirectory(featuresDir, [".md"], {
-    fileLimit: tuning.featuresScan || 10,
-    sortByMtime: true,
-  }, logger);
-  context.features = featureFiles.map((f) => extractFeatureSummary(f.content, f.name));
-
-  // Library
-  const libraryPath = paths.library?.path || "library/";
-  const libraryDir = resolve(wsPath, libraryPath);
-  context.libraryFiles = scanDirectory(libraryDir, [".md"], {
-    fileLimit: 10,
-    contentLimit: tuning.documentSummary || 2000,
-  }, logger);
-
-  // Library sources
-  const sourcesPath = paths.librarySources?.path || "SOURCES.md";
-  context.librarySources = readOptionalFile(resolve(wsPath, sourcesPath));
-
-  // Web files
-  const webPath = paths.web?.path || "src/web/";
-  const webDir = resolve(wsPath, webPath);
-  context.webFiles = scanDirectory(webDir, [".html", ".css", ".js"], {
-    fileLimit: tuning.sourceScan || 10,
-    contentLimit: tuning.sourceContent || 5000,
-    sortByMtime: true,
-    clean: true,
-  }, logger);
-
-  // Contributing guide
-  const contributingPath = paths.contributing?.path || "CONTRIBUTING.md";
-  context.contributing = readOptionalFile(resolve(wsPath, contributingPath), 2000);
-
-  // Package.json
-  context.packageJson = config.packageJson || readOptionalFile(resolve(wsPath, "package.json"), 3000);
-
-  // Config TOML
-  context.configToml = config.configToml || "";
-
-  // Paths
-  context.writablePaths = config.writablePaths || [];
-  context.readOnlyPaths = config.readOnlyPaths || [];
-
-  // Initial test output
-  try {
-    context.testOutput = execSync("npm test 2>&1", { cwd: wsPath, encoding: "utf8", timeout: 120000 });
-  } catch (err) {
-    context.testOutput = `STDOUT:\n${err.stdout || ""}\nSTDERR:\n${err.stderr || ""}`;
-  }
-
-  return context;
-}
-
-/**
- * Fetch GitHub context using the `gh` CLI.
- * Returns null fields when gh is unavailable or data can't be fetched.
- *
- * @param {Object} options
- * @param {number} [options.issueNumber] - Issue number to fetch
- * @param {number} [options.prNumber] - PR number to fetch
- * @param {string} [options.discussionUrl] - Discussion URL to fetch
- * @param {string} [options.workspacePath] - CWD for gh commands
- * @param {Object} [options.logger]
- * @returns {Object} GitHub context
- */
-export function gatherGitHubContext({ issueNumber, prNumber, discussionUrl, workspacePath, logger = defaultLogger } = {}) {
-  const github = { issues: [], issueDetail: null, prDetail: null, discussionDetail: null };
-  const cwd = workspacePath || process.cwd();
-
-  try {
-    // Fetch open issues list
-    const issuesJson = execSync("gh issue list --state open --limit 20 --json number,title,labels,body,createdAt,updatedAt", {
-      cwd,
-      encoding: "utf8",
-      timeout: 30000,
-    });
-    const rawIssues = JSON.parse(issuesJson);
-    github.issues = filterIssues(rawIssues.map((i) => ({
-      number: i.number,
-      title: i.title,
-      body: i.body,
-      labels: i.labels,
-      created_at: i.createdAt,
-      updated_at: i.updatedAt,
-    })));
-  } catch (err) {
-    logger.info(`[context] Could not fetch issues: ${err.message}`);
-  }
-
-  // Fetch specific issue detail
-  if (issueNumber) {
-    try {
-      const issueJson = execSync(`gh issue view ${issueNumber} --json number,title,body,labels,comments,createdAt`, {
-        cwd,
-        encoding: "utf8",
-        timeout: 30000,
-      });
-      github.issueDetail = JSON.parse(issueJson);
-    } catch (err) {
-      logger.info(`[context] Could not fetch issue #${issueNumber}: ${err.message}`);
-    }
-  }
-
-  // Fetch specific PR detail
-  if (prNumber) {
-    try {
-      const prJson = execSync(`gh pr view ${prNumber} --json number,title,body,files,statusCheckRollup`, {
-        cwd,
-        encoding: "utf8",
-        timeout: 30000,
-      });
-      github.prDetail = JSON.parse(prJson);
-    } catch (err) {
-      logger.info(`[context] Could not fetch PR #${prNumber}: ${err.message}`);
-    }
-  }
-
-  // Fetch discussion
-  if (discussionUrl) {
-    try {
-      // Extract discussion number from URL
-      const match = discussionUrl.match(/discussions\/(\d+)/);
-      if (match) {
-        const num = match[1];
-        const discussionJson = execSync(
-          `gh api graphql -f query='{ repository(owner:"{owner}", name:"{repo}") { discussion(number: ${num}) { title body comments(last: 10) { nodes { body author { login } createdAt } } } } }'`,
-          { cwd, encoding: "utf8", timeout: 30000 },
-        );
-        github.discussionDetail = JSON.parse(discussionJson);
-      }
-    } catch (err) {
-      logger.info(`[context] Could not fetch discussion: ${err.message}`);
-    }
-  }
-
-  return github;
-}
-
-/**
- * Build a user prompt for the given agent from available context.
- *
- * Returns an object with the assembled prompt and optional promptBudget metadata.
- * The prompt includes config-driven limits and per-agent refinements that restore
- * the context precision from the old per-task handlers.
- *
- * @param {string} agentName - Agent name (e.g. "agent-iterate")
- * @param {Object} localContext - From gatherLocalContext()
- * @param {Object} [githubContext] - From gatherGitHubContext() (optional)
- * @param {Object} [options]
- * @param {Object} [options.tuning] - Tuning config for limits
- * @param {Object} [options.config] - Full parsed config (for limits injection)
- * @returns {{ prompt: string, promptBudget: Array|null }}
- */
-export function buildUserPrompt(agentName, localContext, githubContext, { tuning, config } = {}) {
-  const needs = AGENT_CONTEXT[agentName] || AGENT_CONTEXT["agent-iterate"];
-  const refinements = AGENT_REFINEMENTS[agentName] || {};
-  const sections = [];
-  const promptBudget = refinements.trackPromptBudget ? [] : null;
-
-  // Target issue — placed prominently when highlightTargetIssue is set
-  if (refinements.highlightTargetIssue && githubContext?.issueDetail) {
-    const issue = githubContext.issueDetail;
-    const issueSection = [
-      `# Target Issue #${issue.number}: ${issue.title}`,
-      issue.body || "(no description)",
-      `Labels: ${(issue.labels || []).map((l) => typeof l === "string" ? l : l.name).join(", ") || "none"}`,
-      "",
-      "**Focus your transformation on resolving this specific issue.**",
-    ];
-    if (issue.comments?.length > 0) {
-      issueSection.push("\n## Comments");
-      for (const c of issue.comments.slice(-5)) {
-        issueSection.push(`**${c.author?.login || "unknown"}**: ${c.body}`);
-      }
-    }
-    const text = issueSection.join("\n");
-    sections.push(text);
-    if (promptBudget) promptBudget.push({ section: "target-issue", size: text.length, files: "1", notes: "" });
-  }
-
-  // Mission
-  if (needs.mission && localContext.mission) {
-    const text = `# Mission\n\n${localContext.mission}`;
-    sections.push(text);
-    if (promptBudget) promptBudget.push({ section: "mission", size: localContext.mission.length, files: "1", notes: "full" });
-  }
-
-  // Current test state — emphasised for fix-code agent
-  if (localContext.testOutput) {
-    const testPreview = localContext.testOutput.substring(0, 4000);
-    if (refinements.emphasizeTestOutput) {
-      sections.push(`# Failing Test Output\n\nThe tests are currently failing. Fix the root cause.\n\n\`\`\`\n${testPreview}\n\`\`\``);
-    } else {
-      sections.push(`# Current Test State\n\n\`\`\`\n${testPreview}\n\`\`\``);
-    }
-  }
-
-  // Source files
-  if (needs.source && localContext.sourceFiles?.length > 0) {
-    const sourceSection = [`# Source Files (${localContext.sourceFiles.length})`];
-    for (const f of localContext.sourceFiles) {
-      sourceSection.push(`## ${f.name}\n\`\`\`\n${f.content}\n\`\`\``);
-    }
-    const text = sourceSection.join("\n\n");
-    sections.push(text);
-    if (promptBudget) promptBudget.push({ section: "source", size: text.length, files: `${localContext.sourceFiles.length}`, notes: "" });
-  }
-
-  // Web files — only when refinements request it
-  if (refinements.includeWebFiles && localContext.webFiles?.length > 0) {
-    const webSection = [`# Website Files (${localContext.webFiles.length})`];
-    for (const f of localContext.webFiles) {
-      webSection.push(`## ${f.name}\n\`\`\`\n${f.content}\n\`\`\``);
-    }
-    const text = webSection.join("\n\n");
-    sections.push(text);
-    if (promptBudget) promptBudget.push({ section: "web", size: text.length, files: `${localContext.webFiles.length}`, notes: "" });
-  }
-
-  // Test files
-  if (needs.tests && localContext.testFiles?.length > 0) {
-    const testSection = [`# Test Files (${localContext.testFiles.length})`];
-    for (const f of localContext.testFiles) {
-      testSection.push(`## ${f.name}\n\`\`\`\n${f.content}\n\`\`\``);
-    }
-    const text = testSection.join("\n\n");
-    sections.push(text);
-    if (promptBudget) promptBudget.push({ section: "tests", size: text.length, files: `${localContext.testFiles.length}`, notes: "" });
-  }
-
-  // Features — sorted incomplete-first when refinements request it
-  if (needs.features && localContext.features?.length > 0) {
-    let features = [...localContext.features];
-    if (refinements.sortFeatures === "incomplete-first") {
-      features.sort((a, b) => {
-        const aIncomplete = /Remaining:/.test(a) || /\[ \]/.test(a) ? 0 : 1;
-        const bIncomplete = /Remaining:/.test(b) || /\[ \]/.test(b) ? 0 : 1;
-        return aIncomplete - bIncomplete;
-      });
-    }
-
-    const limit = config?.paths?.features?.limit;
-    const header = limit
-      ? `# Features (${features.length}/${limit} max)`
-      : `# Features (${features.length})`;
-    const featureSection = [header];
-    for (const f of features) {
-      featureSection.push(f);
-    }
-    const text = featureSection.join("\n\n");
-    sections.push(text);
-    if (promptBudget) promptBudget.push({ section: "features", size: text.length, files: `${features.length}`, notes: "" });
-  }
-
-  // Library
-  if (needs.library && localContext.libraryFiles?.length > 0) {
-    const libLimit = config?.paths?.library?.limit;
-    const header = libLimit
-      ? `# Library Files (${localContext.libraryFiles.length}/${libLimit} max)`
-      : `# Library Files (${localContext.libraryFiles.length})`;
-    const libSection = [header];
-    for (const f of localContext.libraryFiles) {
-      libSection.push(`## ${f.name}\n${f.content}`);
-    }
-    sections.push(libSection.join("\n\n"));
-  }
-
-  // Library sources — vary strategy based on URL presence
-  if (needs.librarySources && localContext.librarySources) {
-    if (refinements.checkSourcesForUrls) {
-      const hasUrls = /https?:\/\//.test(localContext.librarySources);
-      if (!hasUrls) {
-        sections.push(`# Sources\n\n${localContext.librarySources}\n\nPopulate SOURCES.md with 3-8 relevant reference URLs aligned with the mission.`);
-      } else {
-        sections.push(`# Sources\n\n${localContext.librarySources}`);
-      }
-    } else {
-      sections.push(`# Sources\n\n${localContext.librarySources}`);
-    }
-  }
-
-  // Issues (from GitHub context) — not duplicated if target issue already shown
-  if (needs.issues && githubContext?.issues?.length > 0) {
-    const issueSection = [`# Open Issues (${githubContext.issues.length})`];
-    for (const issue of githubContext.issues) {
-      issueSection.push(summariseIssue(issue, tuning?.issueBodyLimit || 500));
-    }
-    const text = issueSection.join("\n\n");
-    sections.push(text);
-    if (promptBudget) promptBudget.push({ section: "issues", size: text.length, files: `${githubContext.issues.length}`, notes: "" });
-  }
-
-  // Specific issue detail (only if not already highlighted as target issue)
-  if (!refinements.highlightTargetIssue && githubContext?.issueDetail) {
-    const issue = githubContext.issueDetail;
-    const issueSection = [`# Issue #${issue.number}: ${issue.title}\n\n${issue.body || "(no body)"}`];
-    if (issue.comments?.length > 0) {
-      issueSection.push("## Comments");
-      for (const c of issue.comments.slice(-10)) {
-        issueSection.push(`**${c.author?.login || "unknown"}**: ${c.body}`);
-      }
-    }
-    sections.push(issueSection.join("\n\n"));
-  }
-
-  // Specific PR detail
-  if (githubContext?.prDetail) {
-    const pr = githubContext.prDetail;
-    const prSection = [`# PR #${pr.number}: ${pr.title}\n\n${pr.body || "(no body)"}`];
-    if (pr.files?.length > 0) {
-      prSection.push(`## Changed Files\n${pr.files.map((f) => `- ${f.path}`).join("\n")}`);
-    }
-    sections.push(prSection.join("\n\n"));
-  }
-
-  // File paths section
-  if (localContext.writablePaths?.length > 0 || localContext.readOnlyPaths?.length > 0) {
-    sections.push(formatPathsSection(
-      localContext.writablePaths || [],
-      localContext.readOnlyPaths || [],
-      { configToml: localContext.configToml, packageJson: localContext.packageJson },
-    ));
-  }
-
-  // Limits section — inject concrete numbers from agentic-lib.toml
-  if (config) {
-    const limitsLines = ["# Limits (from agentic-lib.toml)", ""];
-    const budget = config.transformationBudget || 0;
-    const featLimit = config.paths?.features?.limit;
-    const libLimit = config.paths?.library?.limit;
-
-    if (featLimit) limitsLines.push(`- Maximum feature files: ${featLimit}`);
-    if (libLimit) limitsLines.push(`- Maximum library documents: ${libLimit}`);
-    if (budget > 0) {
-      const intentionPath = config.intentionBot?.intentionFilepath;
-      const cumulativeCost = readCumulativeCost(intentionPath);
-      const remaining = Math.max(0, budget - cumulativeCost);
-      limitsLines.push(`- Transformation budget: ${budget} (used: ${cumulativeCost}, remaining: ${remaining})`);
-    }
-    if (config.featureDevelopmentIssuesWipLimit) limitsLines.push(`- Maximum concurrent feature issues: ${config.featureDevelopmentIssuesWipLimit}`);
-    if (config.maintenanceIssuesWipLimit) limitsLines.push(`- Maximum concurrent maintenance issues: ${config.maintenanceIssuesWipLimit}`);
-    if (config.attemptsPerBranch) limitsLines.push(`- Maximum attempts per branch: ${config.attemptsPerBranch}`);
-    if (config.attemptsPerIssue) limitsLines.push(`- Maximum attempts per issue: ${config.attemptsPerIssue}`);
-
-    if (limitsLines.length > 2) {
-      sections.push(limitsLines.join("\n"));
-    }
-  }
-
-  // Instructions
-  sections.push([
-    "Implement this mission. Read the existing source code and tests,",
-    "make the required changes, run run_tests to verify, and iterate until all tests pass.",
-    "",
-    "Start by reading the existing files, then implement the solution.",
-  ].join("\n"));
-
-  return { prompt: sections.join("\n\n"), promptBudget };
-}