monomind 2.7.12 → 2.7.13

package/packages/@monomind/cli/dist/src/monovector/command-outcomes.js

@@ -1,126 +0,0 @@
-/**
- * Time-windowed command outcome store. Lets post-task derive a MEASURED task
- * success signal from real command exit codes, instead of trusting a
- * caller-supplied --success flag.
- */
-import { promises as fs, statSync } from 'node:fs';
-import { join } from 'node:path';
-/** Refuse to read files larger than this to prevent OOM. */
-const MAX_FILE_BYTES = 50 * 1024 * 1024; // 50 MB
-/** Cap command string length to prevent file bloat. */
-const MAX_COMMAND_LEN = 500;
-function storePath(baseDir) {
-    return join(baseDir, 'command-outcomes.jsonl');
-}
-/** Maximum number of command-outcome records to keep.
- *  deriveRecentSuccess only uses records within a 5-minute window (typically < 50), so
- *  anything older is dead weight. 500 gives a comfortable buffer. */
-const MAX_COMMAND_RECORDS = 500;
-/** Append a command outcome. Non-fatal on error. */
-export async function recordCommand(baseDir, cmd) {
-    try {
-        await fs.mkdir(baseDir, { recursive: true });
-        const path = storePath(baseDir);
-        // Cap command length to prevent individual records from bloating the file.
-        const safeCommand = cmd.command.length > MAX_COMMAND_LEN ? cmd.command.slice(0, MAX_COMMAND_LEN) : cmd.command;
-        const rec = { ts: cmd.ts, command: safeCommand, exitCode: cmd.exitCode, success: cmd.exitCode === 0 };
-        await fs.appendFile(path, JSON.stringify(rec) + '\n', 'utf8');
-        // Opportunistic trim: rewrite only when the file exceeds the cap.
-        // Avoids an extra stat() on every call by catching the overcount lazily.
-        // Guard with size check first to prevent OOM on unexpectedly large files.
-        try {
-            if (statSync(path).size > MAX_FILE_BYTES)
-                return;
-        }
-        catch {
-            return;
-        }
-        const content = await fs.readFile(path, 'utf8').catch(() => '');
-        const lines = content.trim().split('\n').filter(Boolean);
-        if (lines.length > MAX_COMMAND_RECORDS) {
-            await fs.writeFile(path, lines.slice(-MAX_COMMAND_RECORDS).join('\n') + '\n', 'utf8');
-        }
-    }
-    catch (e) {
-        if (process.env.DEBUG || process.env.MONOMIND_DEBUG)
-            console.error('[recordCommand] failed to record command outcome:', e);
-    }
-}
-/**
- * Derive a measured success signal from recent command outcomes.
- * Returns:
- *   - true  if the most recent command(s) in-window ended in a good (final) state
- *   - false if the task ended on a failing command
- *   - null  if there are no recent commands (no signal — caller should treat as unknown)
- *
- * Heuristic: FINAL-STATE — the exit code of the LAST command in the window.
- * Within a task the dominant workflow is iterate-until-green (run tests → fail → fix →
- * run tests → pass), so the task is judged by the state it ENDED in, not by whether any
- * intermediate command failed. "All must pass" would produce pervasive false failures
- * because benign commands routinely exit non-zero (`grep` no-match → 1, `test -f`, `diff`)
- * and the fail-then-pass shape is the norm. Using the last command correctly scores
- * fail→fix→pass as success and ends-on-failure as failure.
- *
- * Known limitation: the store is global (not task-scoped — no taskId exists at command
- * time), so a trailing benign command after a real failure (e.g. `git status` after a
- * failed test) could mask it, and the final command of a prior task could bleed in if a
- * new task records no commands of its own. The time window bounds the latter; per-task
- * scoping would require threading a taskId into post-command.
- */
-export async function deriveRecentSuccess(baseDir, windowMs = 300_000) {
-    try {
-        const p = storePath(baseDir);
-        try {
-            if (statSync(p).size > MAX_FILE_BYTES)
-                return null;
-        }
-        catch { /* file absent */ }
-        const content = await fs.readFile(p, 'utf8').catch(() => '');
-        if (!content)
-            return null;
-        const now = Date.now();
-        const recent = [];
-        for (const line of content.trim().split('\n')) {
-            try {
-                const rec = JSON.parse(line);
-                if (now - rec.ts <= windowMs)
-                    recent.push(rec);
-            }
-            catch { /* skip malformed */ }
-        }
-        if (recent.length === 0)
-            return null;
-        // Final-state: the last command in the window decides the outcome.
-        return recent[recent.length - 1].success;
-    }
-    catch {
-        return null;
-    }
-}
-/** Read recent command outcomes (for diagnostics). */
-export async function readCommandOutcomes(baseDir, windowMs = 300_000) {
-    try {
-        const p = storePath(baseDir);
-        try {
-            if (statSync(p).size > MAX_FILE_BYTES)
-                return [];
-        }
-        catch { /* file absent */ }
-        const content = await fs.readFile(p, 'utf8').catch(() => '');
-        if (!content)
-            return [];
-        const now = Date.now();
-        return content.trim().split('\n').map(l => {
-            try {
-                return JSON.parse(l);
-            }
-            catch {
-                return null;
-            }
-        }).filter((r) => r !== null && now - r.ts <= windowMs);
-    }
-    catch {
-        return [];
-    }
-}
-//# sourceMappingURL=command-outcomes.js.map

package/packages/@monomind/cli/dist/src/monovector/coverage-router.d.ts

@@ -1,103 +0,0 @@
-/**
- * Coverage-Aware Routing (ADR-017)
- *
- * Reads real test-coverage data from disk (Jest/Istanbul `coverage-summary.json`,
- * `lcov.info`, or nyc `out.json`), finds files below a coverage threshold, assigns
- * each gap to an appropriate agent, and produces routing/suggestion decisions.
- *
- * Pure, dependency-light: no native packages, no network. Used by the
- * `monomind route coverage` CLI command and the `coverage_*` MCP tools.
- *
- * @module @monomind/cli/monovector/coverage-router
- */
-/** Per-file coverage percentages (0–100). */
-export interface CoverageEntry {
-    filePath: string;
-    lines: number;
-    branches: number;
-    functions: number;
-    statements: number;
-}
-export interface CoverageSummary {
-    totalFiles: number;
-    overallLineCoverage: number;
-    overallBranchCoverage: number;
-    overallFunctionCoverage: number;
-    overallStatementCoverage: number;
-}
-export interface CoverageData {
-    found: boolean;
-    source: string;
-    entries: CoverageEntry[];
-    summary: CoverageSummary;
-}
-/** A single file below threshold, with its assigned agent. */
-export interface CoverageGap {
-    file: string;
-    currentCoverage: number;
-    gap: number;
-    suggestedAgent: string;
-}
-export interface CoverageGapsResult {
-    found: boolean;
-    totalGaps: number;
-    summary: string;
-    byAgent: Record<string, string[]>;
-    gaps: CoverageGap[];
-}
-export interface CoverageSuggestion {
-    file: string;
-    currentCoverage: number;
-    targetCoverage: number;
-    priority: number;
-    suggestedTests: string[];
-}
-export interface CoverageSuggestResult {
-    found: boolean;
-    path: string;
-    totalGap: number;
-    estimatedEffort: number;
-    suggestions: CoverageSuggestion[];
-}
-export type CoverageAction = 'add-tests' | 'review-coverage' | 'prioritize' | 'skip';
-export interface CoverageRouteResult {
-    found: boolean;
-    action: CoverageAction;
-    priority: number;
-    impactScore: number;
-    estimatedEffort: number;
-    testTypes: string[];
-    targetFiles: string[];
-    gaps: Array<{
-        file: string;
-        currentCoverage: number;
-        gap: number;
-    }>;
-}
-interface CoverageOptions {
-    threshold?: number;
-}
-/**
- * Read coverage data from disk. Checks, in order:
- *  1. coverage/coverage-summary.json  (Jest/Istanbul)
- *  2. coverage-summary.json           (repo root)
- *  3. coverage/lcov.info / lcov.info  (lcov)
- *  4. .nyc_output/out.json            (nyc)
- * Returns `{ found: false }` when no coverage report exists.
- */
-export declare function readCoverage(cwd?: string): CoverageData;
-/** Assign a coverage gap to the most appropriate agent based on file path. */
-export declare function assignAgent(filePath: string): string;
-/** List coverage gaps below threshold, grouped by assigned agent. */
-export declare function coverageGaps(opts?: CoverageOptions & {
-    groupByAgent?: boolean;
-    path?: string;
-}): Promise<CoverageGapsResult>;
-/** Suggest concrete coverage improvements for a path. */
-export declare function coverageSuggest(path?: string, opts?: CoverageOptions & {
-    limit?: number;
-}): Promise<CoverageSuggestResult>;
-/** Produce a coverage-aware routing decision. */
-export declare function coverageRoute(path?: string, opts?: CoverageOptions): Promise<CoverageRouteResult>;
-export {};
-//# sourceMappingURL=coverage-router.d.ts.map

package/packages/@monomind/cli/dist/src/monovector/coverage-router.js

@@ -1,341 +0,0 @@
-/**
- * Coverage-Aware Routing (ADR-017)
- *
- * Reads real test-coverage data from disk (Jest/Istanbul `coverage-summary.json`,
- * `lcov.info`, or nyc `out.json`), finds files below a coverage threshold, assigns
- * each gap to an appropriate agent, and produces routing/suggestion decisions.
- *
- * Pure, dependency-light: no native packages, no network. Used by the
- * `monomind route coverage` CLI command and the `coverage_*` MCP tools.
- *
- * @module @monomind/cli/monovector/coverage-router
- */
-import { existsSync, readFileSync, statSync } from 'fs';
-import { join } from 'path';
-import { getProjectCwd } from '../mcp-tools/types.js';
-// ============================================================================
-// Coverage reader (Jest/Istanbul summary, lcov, nyc)
-// ============================================================================
-const EMPTY = {
-    found: false,
-    source: 'none',
-    entries: [],
-    summary: {
-        totalFiles: 0,
-        overallLineCoverage: 0,
-        overallBranchCoverage: 0,
-        overallFunctionCoverage: 0,
-        overallStatementCoverage: 0,
-    },
-};
-/** Maximum bytes for a JSON coverage summary (20 MB). */
-const MAX_COVERAGE_JSON_BYTES = 20 * 1024 * 1024;
-/** Maximum bytes for an lcov.info file (50 MB). */
-const MAX_COVERAGE_LCOV_BYTES = 50 * 1024 * 1024;
-/**
- * Read coverage data from disk. Checks, in order:
- *  1. coverage/coverage-summary.json  (Jest/Istanbul)
- *  2. coverage-summary.json           (repo root)
- *  3. coverage/lcov.info / lcov.info  (lcov)
- *  4. .nyc_output/out.json            (nyc)
- * Returns `{ found: false }` when no coverage report exists.
- */
-export function readCoverage(cwd = getProjectCwd()) {
-    for (const rel of ['coverage/coverage-summary.json', 'coverage-summary.json']) {
-        const p = join(cwd, rel);
-        if (existsSync(p) && statSync(p).size <= MAX_COVERAGE_JSON_BYTES) {
-            try {
-                return parseSummaryJson(JSON.parse(readFileSync(p, 'utf-8')), rel);
-            }
-            catch {
-                /* malformed — try next */
-            }
-        }
-    }
-    for (const rel of ['coverage/lcov.info', 'lcov.info']) {
-        const p = join(cwd, rel);
-        if (existsSync(p) && statSync(p).size <= MAX_COVERAGE_LCOV_BYTES) {
-            try {
-                return parseLcov(readFileSync(p, 'utf-8'), rel);
-            }
-            catch {
-                /* malformed — try next */
-            }
-        }
-    }
-    const nyc = join(cwd, '.nyc_output', 'out.json');
-    if (existsSync(nyc) && statSync(nyc).size <= MAX_COVERAGE_JSON_BYTES) {
-        try {
-            return parseSummaryJson(JSON.parse(readFileSync(nyc, 'utf-8')), '.nyc_output/out.json');
-        }
-        catch {
-            /* malformed */
-        }
-    }
-    return EMPTY;
-}
-function pct(covered, total, fallback = 100) {
-    if (total == null || total === 0)
-        return fallback;
-    return ((covered ?? 0) / total) * 100;
-}
-function parseSummaryJson(data, source) {
-    const entries = [];
-    let tl = 0, cl = 0, tb = 0, cb = 0, tf = 0, cf = 0, ts = 0, cs = 0;
-    for (const [filePath, metrics] of Object.entries(data)) {
-        if (filePath === 'total')
-            continue;
-        const m = metrics;
-        if (!m || typeof m !== 'object')
-            continue;
-        entries.push({
-            filePath,
-            lines: m.lines?.pct ?? pct(m.lines?.covered, m.lines?.total, 0),
-            branches: m.branches?.pct ?? pct(m.branches?.covered, m.branches?.total),
-            functions: m.functions?.pct ?? pct(m.functions?.covered, m.functions?.total),
-            statements: m.statements?.pct ?? pct(m.statements?.covered, m.statements?.total),
-        });
-        tl += m.lines?.total ?? 0;
-        cl += m.lines?.covered ?? 0;
-        tb += m.branches?.total ?? 0;
-        cb += m.branches?.covered ?? 0;
-        tf += m.functions?.total ?? 0;
-        cf += m.functions?.covered ?? 0;
-        ts += m.statements?.total ?? 0;
-        cs += m.statements?.covered ?? 0;
-    }
-    const total = data['total'];
-    entries.sort((a, b) => a.lines - b.lines);
-    return {
-        found: true,
-        source,
-        entries,
-        summary: {
-            totalFiles: entries.length,
-            overallLineCoverage: total?.lines?.pct ?? pct(cl, tl, 0),
-            overallBranchCoverage: total?.branches?.pct ?? pct(cb, tb, 0),
-            overallFunctionCoverage: total?.functions?.pct ?? pct(cf, tf, 0),
-            overallStatementCoverage: total?.statements?.pct ?? pct(cs, ts, 0),
-        },
-    };
-}
-function parseLcov(raw, source) {
-    const entries = [];
-    let file = '';
-    let lf = 0, lh = 0, brf = 0, brh = 0, fnf = 0, fnh = 0;
-    let tl = 0, cl = 0, tb = 0, cb = 0, tf = 0, cf = 0;
-    const flush = () => {
-        if (!file)
-            return;
-        entries.push({
-            filePath: file,
-            lines: pct(lh, lf, 0),
-            branches: pct(brh, brf),
-            functions: pct(fnh, fnf),
-            statements: pct(lh, lf, 0),
-        });
-        tl += lf;
-        cl += lh;
-        tb += brf;
-        cb += brh;
-        tf += fnf;
-        cf += fnh;
-        file = '';
-        lf = lh = brf = brh = fnf = fnh = 0;
-    };
-    for (const line of raw.split('\n')) {
-        const t = line.trim();
-        if (t.startsWith('SF:')) {
-            flush();
-            file = t.slice(3);
-        }
-        else if (t.startsWith('LF:'))
-            lf = Number(t.slice(3)) || 0;
-        else if (t.startsWith('LH:'))
-            lh = Number(t.slice(3)) || 0;
-        else if (t.startsWith('BRF:'))
-            brf = Number(t.slice(4)) || 0;
-        else if (t.startsWith('BRH:'))
-            brh = Number(t.slice(4)) || 0;
-        else if (t.startsWith('FNF:'))
-            fnf = Number(t.slice(4)) || 0;
-        else if (t.startsWith('FNH:'))
-            fnh = Number(t.slice(4)) || 0;
-        else if (t === 'end_of_record')
-            flush();
-    }
-    flush();
-    entries.sort((a, b) => a.lines - b.lines);
-    return {
-        found: entries.length > 0,
-        source,
-        entries,
-        summary: {
-            totalFiles: entries.length,
-            overallLineCoverage: pct(cl, tl, 0),
-            overallBranchCoverage: pct(cb, tb, 0),
-            overallFunctionCoverage: pct(cf, tf, 0),
-            overallStatementCoverage: pct(cl, tl, 0),
-        },
-    };
-}
-// ============================================================================
-// Agent assignment + heuristics
-// ============================================================================
-/** Assign a coverage gap to the most appropriate agent based on file path. */
-export function assignAgent(filePath) {
-    const p = filePath.toLowerCase();
-    if (/(security|auth|crypto|password|token|permission|sanitiz)/.test(p))
-        return 'security-architect';
-    if (/(api|route|controller|endpoint|handler|server|http)/.test(p))
-        return 'backend-dev';
-    if (/(\.tsx|\.jsx|component|ui\/|view|page|render)/.test(p))
-        return 'frontend-developer';
-    if (/(db|database|migration|schema|model|repository|query)/.test(p))
-        return 'backend-dev';
-    if (/(util|helper|lib\/|common)/.test(p))
-        return 'coder';
-    return 'tester';
-}
-/** Suggest test types based on which coverage dimensions are weakest. */
-function suggestTestTypes(entry) {
-    const types = [];
-    if (entry.lines < 80 || entry.statements < 80)
-        types.push('unit');
-    if (entry.branches < 80)
-        types.push('edge-case');
-    if (entry.functions < 80)
-        types.push('unit');
-    if (entry.lines < 50)
-        types.push('integration');
-    return Array.from(new Set(types.length ? types : ['unit']));
-}
-/** Concrete test suggestions for a file below threshold. */
-function suggestTests(entry, target) {
-    const out = [];
-    const base = entry.filePath.replace(/\.[tj]sx?$/, '');
-    if (entry.functions < target)
-        out.push(`Add unit tests covering untested functions in ${entry.filePath}`);
-    if (entry.branches < target)
-        out.push(`Add tests for uncovered branches/conditionals (currently ${entry.branches.toFixed(0)}%)`);
-    if (entry.lines < 50)
-        out.push(`Add an integration test exercising the main path of ${entry.filePath}`);
-    out.push(`Create ${base}.test.ts with edge-case and error-path coverage`);
-    return out.slice(0, 4);
-}
-/** Effort estimate (hours) to close a coverage gap on a file. */
-function estimateEffort(gap) {
-    // ~0.5h per 10% gap, floored at 0.25h, capped at 4h per file.
-    return Math.min(4, Math.max(0.25, (gap / 10) * 0.5));
-}
-function filterByPath(entries, path) {
-    if (!path || path === '.' || path === './')
-        return entries;
-    const norm = path.replace(/^\.\//, '');
-    return entries.filter(e => e.filePath.includes(norm));
-}
-// ============================================================================
-// Public API (consumed by route.ts and coverage-tools.ts)
-// ============================================================================
-/** List coverage gaps below threshold, grouped by assigned agent. */
-export async function coverageGaps(opts = {}) {
-    const threshold = opts.threshold ?? 80;
-    const data = readCoverage();
-    if (!data.found) {
-        return { found: false, totalGaps: 0, summary: 'No coverage report found. Run your test suite with coverage enabled.', byAgent: {}, gaps: [] };
-    }
-    const entries = filterByPath(data.entries, opts.path ?? '');
-    const gaps = entries
-        .filter(e => e.lines < threshold)
-        .map(e => ({
-        file: e.filePath,
-        currentCoverage: e.lines,
-        gap: threshold - e.lines,
-        suggestedAgent: assignAgent(e.filePath),
-    }))
-        .sort((a, b) => b.gap - a.gap);
-    const byAgent = {};
-    for (const g of gaps)
-        (byAgent[g.suggestedAgent] ??= []).push(g.file);
-    return {
-        found: true,
-        totalGaps: gaps.length,
-        summary: gaps.length === 0
-            ? `All ${entries.length} files meet the ${threshold}% threshold.`
-            : `${gaps.length} of ${entries.length} files below ${threshold}% (overall ${data.summary.overallLineCoverage.toFixed(1)}%).`,
-        byAgent,
-        gaps,
-    };
-}
-/** Suggest concrete coverage improvements for a path. */
-export async function coverageSuggest(path = '.', opts = {}) {
-    const threshold = opts.threshold ?? 80;
-    const limit = opts.limit ?? 20;
-    const data = readCoverage();
-    if (!data.found) {
-        return { found: false, path, totalGap: 0, estimatedEffort: 0, suggestions: [] };
-    }
-    const entries = filterByPath(data.entries, path).filter(e => e.lines < threshold);
-    const suggestions = entries
-        .map(e => {
-        const gap = threshold - e.lines;
-        // priority 1–10 from gap size (a 40%+ gap → 10).
-        const priority = Math.min(10, Math.max(1, Math.round(gap / 4)));
-        return {
-            file: e.filePath,
-            currentCoverage: e.lines,
-            targetCoverage: threshold,
-            priority,
-            suggestedTests: suggestTests(e, threshold),
-        };
-    })
-        .sort((a, b) => b.priority - a.priority)
-        .slice(0, limit);
-    const totalGap = entries.reduce((s, e) => s + (threshold - e.lines), 0);
-    const estimatedEffort = entries.reduce((s, e) => s + estimateEffort(threshold - e.lines), 0);
-    return { found: true, path, totalGap, estimatedEffort, suggestions };
-}
-/** Produce a coverage-aware routing decision. */
-export async function coverageRoute(path = '', opts = {}) {
-    const threshold = opts.threshold ?? 80;
-    const data = readCoverage();
-    if (!data.found) {
-        return { found: false, action: 'skip', priority: 0, impactScore: 0, estimatedEffort: 0, testTypes: [], targetFiles: [], gaps: [] };
-    }
-    const entries = filterByPath(data.entries, path);
-    const below = entries.filter(e => e.lines < threshold);
-    const overall = data.summary.overallLineCoverage;
-    const gaps = below
-        .map(e => ({ file: e.filePath, currentCoverage: e.lines, gap: threshold - e.lines }))
-        .sort((a, b) => b.gap - a.gap);
-    let action;
-    if (below.length === 0)
-        action = 'skip';
-    else if (overall < threshold - 20)
-        action = 'prioritize';
-    else if (overall < threshold)
-        action = 'add-tests';
-    else
-        action = 'review-coverage';
-    // priority 1–10: severity of the worst gap + breadth.
-    const worstGap = gaps[0]?.gap ?? 0;
-    const breadth = entries.length ? below.length / entries.length : 0;
-    const priority = action === 'skip' ? 0 : Math.min(10, Math.max(1, Math.round((worstGap / 5) + breadth * 4)));
-    // impactScore 0–100: how far overall coverage sits below threshold — i.e. the
-    // headroom (in coverage points) that closing these gaps would reclaim.
-    const impactScore = Math.round(Math.min(100, Math.max(0, threshold - overall)));
-    const estimatedEffort = Number(below.reduce((s, e) => s + estimateEffort(threshold - e.lines), 0).toFixed(1));
-    // aggregate test types across the worst files
-    const testTypes = Array.from(new Set(below.slice(0, 10).flatMap(suggestTestTypes)));
-    return {
-        found: true,
-        action,
-        priority,
-        impactScore,
-        estimatedEffort,
-        testTypes: testTypes.length ? testTypes : ['unit'],
-        targetFiles: gaps.map(g => g.file),
-        gaps,
-    };
-}
-//# sourceMappingURL=coverage-router.js.map

package/packages/@monomind/cli/dist/src/monovector/coverage-tools.d.ts

@@ -1,14 +0,0 @@
-/**
- * Coverage Router MCP Tools (ADR-017)
- *
- * Exposes coverage-aware routing over MCP: analyze test-coverage data, list gaps
- * with agent assignments, and suggest concrete test improvements. Thin wrappers
- * around `./coverage-router.js` (the shared pure logic also used by the
- * `monomind route coverage` CLI command).
- *
- * @module @monomind/cli/monovector/coverage-tools
- */
-import type { MCPTool } from '../mcp-tools/types.js';
-export declare const coverageRouterTools: MCPTool[];
-export default coverageRouterTools;
-//# sourceMappingURL=coverage-tools.d.ts.map