@selvakumaresra/specship 0.1.0

package/scripts/agent-eval/parse-arms.mjs

@@ -0,0 +1,116 @@
+#!/usr/bin/env node
+// Analyze the tool-surface ablation (/tmp/arms/<repo>/<arm>-r<n>.jsonl).
+// Compares arms A–E on trace adoption, Read/Grep fallback, specship payload,
+// round-trips, and duration — averaged across runs per arm.
+//
+// The decisive signal is READS: if removing a tool raises Reads on a question
+// class, that tool was load-bearing for it (not redundant). If removing it
+// changes nothing, it was redundant.
+//
+//   A control       all tools            no steering   (baseline)
+//   B steer         all tools            trace-first   (adoption)
+//   C no-explore    hide explore         trace-first   (is explore redundant?)
+//   D trace-centric hide explore+context trace-first   (is the survey pair redundant?)
+//   E control-probe hide explore+context trace-first   (NON-flow Q — should degrade)
+//
+// Usage: node scripts/agent-eval/parse-arms.mjs [/tmp/arms]
+import { readFileSync, readdirSync, existsSync, statSync } from 'fs';
+import { join } from 'path';
+
+const ROOT = process.argv[2] || '/tmp/arms';
+const cgShort = (n) => n.replace('mcp__specship__specship_', '').replace('mcp__specship__', '');
+
+function parse(file) {
+  if (!existsSync(file)) return null;
+  const lines = readFileSync(file, 'utf8').split('\n').filter(Boolean);
+  const calls = []; let result = null, initCg = 0;
+  for (const l of lines) {
+    let ev; try { ev = JSON.parse(l); } catch { continue; }
+    if (ev.type === 'system' && ev.subtype === 'init') initCg = (ev.tools || []).filter(t => /specship/.test(t)).length;
+    if (ev.type === 'assistant') for (const b of (ev.message?.content || [])) if (b.type === 'tool_use')
+      calls.push({ id: b.id, name: b.name, out: 0 });
+    if (ev.type === 'user') for (const b of (ev.message?.content || [])) if (b.type === 'tool_result') {
+      const c = b.content;
+      const txt = typeof c === 'string' ? c : Array.isArray(c) ? c.map(x => x?.text || '').join('') : '';
+      const call = calls.find(k => k.id === b.tool_use_id); if (call) call.out = txt.length;
+    }
+    if (ev.type === 'result') result = ev;
+  }
+  const cg = calls.filter(c => c.name.includes('specship'));
+  return {
+    initCg,
+    reads: calls.filter(c => c.name === 'Read').length,
+    greps: calls.filter(c => c.name === 'Grep').length + calls.filter(c => c.name === 'Glob').length,
+    cgCalls: cg.length,
+    cgSeq: cg.map(c => cgShort(c.name)),
+    cgOut: cg.reduce((s, c) => s + c.out, 0),
+    traceUsed: cg.some(c => c.name.includes('trace')),
+    turns: result?.num_turns ?? null,
+    dur: result?.duration_ms ? Math.round(result.duration_ms / 1000) : null,
+    cost: result?.total_cost_usd || 0,
+    ok: result?.subtype === 'success',
+  };
+}
+
+// repo -> arm -> [runs]
+const data = {};
+if (!existsSync(ROOT)) { console.error(`no ${ROOT}`); process.exit(1); }
+for (const repo of readdirSync(ROOT)) {
+  const rdir = join(ROOT, repo);
+  if (!statSync(rdir).isDirectory()) continue;
+  for (const f of readdirSync(rdir)) {
+    const m = f.match(/^([A-I])-r(\d+)\.jsonl$/); if (!m) continue;
+    const p = parse(join(rdir, f)); if (!p || !p.ok) continue;
+    (((data[repo] ??= {})[m[1]]) ??= []).push(p);
+  }
+}
+
+const avg = (a, f) => a.length ? a.reduce((s, x) => s + (f(x) || 0), 0) / a.length : 0;
+const k = (n) => (n / 1000).toFixed(1);
+const pad = (s, n) => String(s).padEnd(n);
+const ARMS = ['A', 'H', 'I', 'B', 'F', 'G', 'C', 'D', 'E'];
+const LABEL = { A: 'A all/none(old)', H: 'H body-trace/none', I: 'I bodytrace+dest', B: 'B all/steer(thin)', F: 'F all/steer(body)', G: 'G ported(noprompt)', C: 'C no-explore', D: 'D trace-centric', E: 'E nonflow-probe' };
+
+// ---- per repo × arm ----
+console.log('\n=== PER REPO × ARM (avg over runs) ===');
+console.log(pad('repo', 22), pad('arm', 16), 'tools', 'trace', pad('reads', 6), pad('cgOutK', 7), pad('turns', 6), 'dur');
+for (const repo of Object.keys(data).sort()) {
+  for (const arm of ARMS) {
+    const runs = data[repo][arm]; if (!runs?.length) continue;
+    console.log(
+      pad(repo, 22), pad(LABEL[arm], 16),
+      pad(runs[0].initCg, 5),
+      pad(runs.filter(r => r.traceUsed).length + '/' + runs.length, 5),
+      pad(avg(runs, r => r.reads).toFixed(1), 6),
+      pad(k(avg(runs, r => r.cgOut)), 7),
+      pad(avg(runs, r => r.turns).toFixed(1), 6),
+      avg(runs, r => r.dur).toFixed(0) + 's',
+    );
+  }
+}
+
+// ---- aggregate per arm (flow arms A–D over the flow repos; E shown apart) ----
+console.log('\n=== AGGREGATE PER ARM (mean across repos) ===');
+console.log(pad('arm', 16), pad('adoption', 9), pad('reads', 7), pad('greps', 7), pad('cgOutK', 8), pad('turns', 7), pad('dur', 6), 'cost');
+for (const arm of ARMS) {
+  const all = [];
+  for (const repo of Object.keys(data)) for (const r of (data[repo][arm] || [])) all.push({ ...r, repo });
+  if (!all.length) continue;
+  const repos = new Set(all.map(r => r.repo)).size;
+  const adopt = all.filter(r => r.traceUsed).length;
+  console.log(
+    pad(LABEL[arm], 16),
+    pad(`${adopt}/${all.length}`, 9),
+    pad(avg(all, r => r.reads).toFixed(2), 7),
+    pad(avg(all, r => r.greps).toFixed(2), 7),
+    pad(k(avg(all, r => r.cgOut)), 8),
+    pad(avg(all, r => r.turns).toFixed(1), 7),
+    pad(avg(all, r => r.dur).toFixed(0) + 's', 6),
+    '$' + avg(all, r => r.cost).toFixed(3),
+    `  (${repos} repos)`,
+  );
+}
+
+console.log('\nRead the signal: B vs A = does steering alone fix adoption + cut payload.');
+console.log('C vs B = is explore redundant (reads should NOT jump). D vs C = is context redundant.');
+console.log('E = non-flow under trace-centric; reads SHOULD jump (proves survey tools are load-bearing).');

package/scripts/agent-eval/parse-bench-readme.mjs

@@ -0,0 +1,84 @@
+#!/usr/bin/env node
+// Aggregate the README A/B (bench-readme.sh output): per repo, median of N runs
+// per arm → time, tool calls, tokens, cost, and % saved. Plus an average row.
+//
+// Tokens = SUM of per-turn assistant `usage` (input + output + cache read +
+// cache creation) — the cumulative "total tokens processed". NOTE: `result.usage`
+// is last-turn-only in current Claude Code, so it under-counts badly; don't use it.
+// `total_cost_usd` and `duration_ms` are already cumulative.
+//
+// Usage: node parse-bench-readme.mjs [/tmp/ab-readme]
+import { readFileSync, existsSync, readdirSync } from 'fs';
+import { join } from 'path';
+const ROOT = process.argv[2] || '/tmp/ab-readme';
+const REPOS = ['vscode', 'excalidraw', 'django', 'tokio', 'okhttp', 'gin', 'alamofire'];
+
+function parse(file) {
+  if (!existsSync(file)) return null;
+  const L = readFileSync(file, 'utf8').split('\n').filter(Boolean);
+  let tools = 0, reads = 0, grep = 0, cg = 0, tokens = 0, r = null, raced = false;
+  for (const l of L) { let e; try { e = JSON.parse(l); } catch { continue; }
+    if (e.type === 'assistant') {
+      const u = e.message?.usage;
+      if (u) tokens += (u.input_tokens || 0) + (u.output_tokens || 0) + (u.cache_read_input_tokens || 0) + (u.cache_creation_input_tokens || 0);
+      for (const b of (e.message?.content || [])) if (b.type === 'tool_use') {
+        const n = b.name;
+        if (n === 'ToolSearch') continue;
+        tools++;
+        if (n === 'Read') reads++;
+        else if (n === 'Grep' || n === 'Glob') grep++;
+        else if (/specship/.test(n)) cg++;
+      }
+    }
+    // MCP cold-start race: the headless agent fired before `specship serve --mcp`
+    // finished registering its tools, so early calls returned "No such tool
+    // available" and the agent floundered into grep/Read. That measures SpecShip's
+    // startup latency, NOT its steady-state value — flag the run so the aggregate
+    // can exclude it (an artifact of headless first-turn timing, not the tool).
+    if (e.type === 'user') for (const b of (Array.isArray(e.message?.content) ? e.message.content : [])) {
+      if (b.type === 'tool_result') {
+        const t = Array.isArray(b.content) ? b.content.map(c => c.text || '').join('') : (b.content || '');
+        if (/No such tool available/.test(t)) raced = true;
+      }
+    }
+    if (e.type === 'result') r = e;
+  }
+  if (!r || r.subtype !== 'success') return null;
+  return { dur: r.duration_ms / 1000, tools, reads, grep, cg, tokens, cost: r.total_cost_usd || 0, raced };
+}
+const median = (arr) => { const v = [...arr].sort((a, b) => a - b); const n = v.length; return n === 0 ? 0 : n % 2 ? v[(n - 1) / 2] : (v[n / 2 - 1] + v[n / 2]) / 2; };
+const fmtTime = (s) => s >= 60 ? `${Math.floor(s / 60)}m ${Math.round(s % 60)}s` : `${Math.round(s)}s`;
+const fmtTok = (t) => t >= 1e6 ? `${(t / 1e6).toFixed(1)}M` : `${Math.round(t / 1000)}k`;
+const pct = (w, wo) => wo > 0 ? Math.round((1 - w / wo) * 100) : 0;
+
+console.log('repo        n(w/wo)  time WITH→WITHOUT      tools W→WO   tokens W→WO (saved)     cost W→WO (saved)');
+const savings = { cost: [], tokens: [], time: [], tools: [] };
+for (const repo of REPOS) {
+  const dir = join(ROOT, repo);
+  const runDirs = existsSync(dir) ? readdirSync(dir).filter(d => /^run\d+$/.test(d)) : [];
+  // Exclude MCP-cold-start-raced WITH runs by default — they measure a startup
+  // race, not steady-state value. `CG_INCLUDE_RACED=1` keeps them (to see the raw
+  // distribution). The WITHOUT arm has no MCP, so it's never raced.
+  const includeRaced = process.env.CG_INCLUDE_RACED === '1';
+  const W = [], WO = []; let racedExcluded = 0;
+  for (const rd of runDirs) {
+    const w = parse(join(dir, rd, 'run-headless-with.jsonl'));
+    if (w) { if (w.raced && !includeRaced) racedExcluded++; else W.push(w); }
+    const wo = parse(join(dir, rd, 'run-headless-without.jsonl')); if (wo) WO.push(wo);
+  }
+  if (!W.length || !WO.length) { console.log(`${repo.padEnd(11)} (incomplete: w=${W.length} wo=${WO.length})`); continue; }
+  const m = (arr, k) => median(arr.map(x => x[k]));
+  const wT = m(W, 'dur'), woT = m(WO, 'dur'), wTok = m(W, 'tokens'), woTok = m(WO, 'tokens');
+  const wC = m(W, 'cost'), woC = m(WO, 'cost'), wTl = m(W, 'tools'), woTl = m(WO, 'tools');
+  savings.time.push(pct(wT, woT)); savings.tokens.push(pct(wTok, woTok)); savings.cost.push(pct(wC, woC)); savings.tools.push(pct(wTl, woTl));
+  console.log(
+    `${repo.padEnd(11)} ${W.length}/${WO.length}      ` +
+    `${(fmtTime(wT) + '→' + fmtTime(woT)).padEnd(22)}` +
+    `${(Math.round(wTl) + '→' + Math.round(woTl)).padEnd(12)}` +
+    `${(fmtTok(wTok) + '→' + fmtTok(woTok) + ' (' + pct(wTok, woTok) + '%)').padEnd(24)}` +
+    `$${wC.toFixed(2)}→$${woC.toFixed(2)} (${pct(wC, woC)}%)` +
+    (racedExcluded ? `  [${racedExcluded} raced run${racedExcluded === 1 ? '' : 's'} excluded]` : '')
+  );
+}
+const avg = (a) => a.length ? Math.round(a.reduce((s, x) => s + x, 0) / a.length) : 0;
+console.log(`\nAVERAGE saved:  cost ${avg(savings.cost)}%  ·  tokens ${avg(savings.tokens)}%  ·  time ${avg(savings.time)}%  ·  tool calls ${avg(savings.tools)}%`);

package/scripts/agent-eval/parse-run.mjs

@@ -0,0 +1,45 @@
+#!/usr/bin/env node
+// Parse a Claude Code stream-json run log: tool-call sequence + token usage.
+import { readFileSync } from 'fs';
+const file = process.argv[2];
+const lines = readFileSync(file, 'utf8').split('\n').filter(Boolean);
+
+const toolCalls = [];
+let result = null;
+let initTools = null;
+
+for (const line of lines) {
+  let ev;
+  try { ev = JSON.parse(line); } catch { continue; }
+  if (ev.type === 'system' && ev.subtype === 'init') {
+    initTools = (ev.tools || []).filter(t => /specship/.test(t));
+  }
+  if (ev.type === 'assistant' && ev.message?.content) {
+    for (const block of ev.message.content) {
+      if (block.type === 'tool_use') {
+        let detail = '';
+        if (block.name === 'Task') detail = ` [subagent_type=${block.input?.subagent_type ?? '?'}] ${(block.input?.description ?? '').slice(0,40)}`;
+        else if (/specship/.test(block.name)) detail = ` ${JSON.stringify(block.input?.query ?? block.input?.task ?? block.input?.symbol ?? '').slice(0,60)}`;
+        else if (block.name === 'Bash') detail = ` ${(block.input?.command ?? '').slice(0,50)}`;
+        else if (block.name === 'Read') detail = ` ${(block.input?.file_path ?? '').split('/').slice(-1)[0]}`;
+        toolCalls.push(`${block.name}${detail}`);
+      }
+    }
+  }
+  if (ev.type === 'result') result = ev;
+}
+
+console.log(`\n=== ${file.split('/').pop()} ===`);
+console.log(`specship tools exposed: ${initTools ? initTools.length : '?'}`);
+console.log(`\nTool calls (${toolCalls.length}):`);
+const counts = {};
+for (const tc of toolCalls) { const n = tc.split(' ')[0]; counts[n] = (counts[n]||0)+1; }
+console.log('  by type:', JSON.stringify(counts));
+toolCalls.forEach((tc, i) => console.log(`  ${i+1}. ${tc}`));
+
+if (result) {
+  const u = result.usage || {};
+  const totalIn = (u.input_tokens||0) + (u.cache_read_input_tokens||0) + (u.cache_creation_input_tokens||0);
+  console.log(`\nResult: ${result.subtype} | duration ${(result.duration_ms/1000).toFixed(0)}s | turns ${result.num_turns}`);
+  console.log(`  tokens: in=${totalIn} out=${u.output_tokens||0} | cost $${(result.total_cost_usd||0).toFixed(3)}`);
+}

package/scripts/agent-eval/parse-session.mjs

@@ -0,0 +1,93 @@
+#!/usr/bin/env node
+// Parse the newest Claude Code session log for a project + its subagent logs,
+// and report the tool-call breakdown (main + subagents). Works for interactive
+// runs (driven via itrun.sh) — Claude Code writes full transcripts to
+// ~/.claude/projects/<escaped-cwd>/<session>.jsonl with subagents/ alongside.
+import { readFileSync, readdirSync, statSync, existsSync, realpathSync } from 'fs';
+import { join } from 'path';
+import { homedir } from 'os';
+
+const projectArg = process.argv[2];
+if (!projectArg) { console.error('usage: parse-session.mjs <project-dir>'); process.exit(1); }
+
+// Claude Code escapes the (real) cwd by replacing every "/" with "-".
+const real = realpathSync(projectArg);
+const escaped = real.replace(/\//g, '-');
+const projDir = join(homedir(), '.claude', 'projects', escaped);
+if (!existsSync(projDir)) { console.error('no session logs at', projDir); process.exit(1); }
+
+// Newest top-level session .jsonl
+const sessions = readdirSync(projDir)
+  .filter(f => f.endsWith('.jsonl'))
+  .map(f => ({ f, m: statSync(join(projDir, f)).mtimeMs }))
+  .sort((a, b) => b.m - a.m);
+if (sessions.length === 0) { console.error('no .jsonl sessions in', projDir); process.exit(1); }
+const sessionId = sessions[0].f.replace('.jsonl', '');
+
+function tally(file) {
+  const counts = {};
+  for (const line of readFileSync(file, 'utf8').split('\n')) {
+    if (!line) continue;
+    let ev; try { ev = JSON.parse(line); } catch { continue; }
+    const content = ev.message?.content;
+    if (!Array.isArray(content)) continue;
+    for (const b of content) {
+      if (b.type === 'tool_use') counts[b.name] = (counts[b.name] || 0) + 1;
+    }
+  }
+  return counts;
+}
+
+// Sum token usage from a transcript. The TUI's "Done (…Xk tokens…)" line only
+// covers a subagent's throughput; this works for main-thread runs too and is
+// consistent across both paths. `gen` = output, `fresh` = uncached input
+// (input + cache_creation), `cached` = cache reads (≈free), `total` = all.
+function sumTokens(file) {
+  const t = { gen: 0, fresh: 0, cached: 0 };
+  for (const line of readFileSync(file, 'utf8').split('\n')) {
+    if (!line) continue;
+    let ev; try { ev = JSON.parse(line); } catch { continue; }
+    const u = ev.message?.usage;
+    if (!u) continue;
+    t.gen += u.output_tokens || 0;
+    t.fresh += (u.input_tokens || 0) + (u.cache_creation_input_tokens || 0);
+    t.cached += u.cache_read_input_tokens || 0;
+  }
+  return t;
+}
+
+const mainCounts = tally(join(projDir, sessionId + '.jsonl'));
+
+// Subagent transcripts live under <session>/subagents/*.jsonl
+const subDir = join(projDir, sessionId, 'subagents');
+const subCounts = {};
+let subAgentFiles = 0;
+if (existsSync(subDir)) {
+  for (const f of readdirSync(subDir).filter(f => f.endsWith('.jsonl'))) {
+    subAgentFiles++;
+    const c = tally(join(subDir, f));
+    for (const [k, v] of Object.entries(c)) subCounts[k] = (subCounts[k] || 0) + v;
+  }
+}
+
+const fmt = (counts) => Object.entries(counts).sort((a, b) => b[1] - a[1])
+  .map(([k, v]) => `    ${String(v).padStart(3)}  ${k}`).join('\n') || '    (none)';
+
+console.log(`session: ${sessionId}`);
+console.log(`\nMAIN thread tools:\n${fmt(mainCounts)}`);
+console.log(`\nSUBAGENT tools (${subAgentFiles} subagent transcript${subAgentFiles === 1 ? '' : 's'}):\n${fmt(subCounts)}`);
+
+const explore = subCounts['mcp__specship__specship_explore'] || mainCounts['mcp__specship__specship_explore'] || 0;
+const reads = (subCounts['Read'] || 0) + (mainCounts['Read'] || 0);
+const greps = (subCounts['Grep'] || 0) + (mainCounts['Grep'] || 0) + (subCounts['Bash'] || 0) + (mainCounts['Bash'] || 0);
+console.log(`\nVERDICT: specship_explore used ${explore}x | Read ${reads} | Grep/Bash ${greps}`);
+
+// Token totals (main + subagents), consistent across main-thread and subagent runs.
+const tok = { gen: 0, fresh: 0, cached: 0 };
+const addTok = (t) => { tok.gen += t.gen; tok.fresh += t.fresh; tok.cached += t.cached; };
+addTok(sumTokens(join(projDir, sessionId + '.jsonl')));
+if (existsSync(subDir)) {
+  for (const f of readdirSync(subDir).filter(f => f.endsWith('.jsonl'))) addTok(sumTokens(join(subDir, f)));
+}
+const k = (n) => (n / 1000).toFixed(1) + 'k';
+console.log(`TOKENS: gen ${k(tok.gen)} | fresh-in ${k(tok.fresh)} | cached-in ${k(tok.cached)} | billable≈ ${k(tok.gen + tok.fresh)}`);

package/scripts/agent-eval/probe-context.mjs

@@ -0,0 +1,21 @@
+#!/usr/bin/env node
+// Probe specship_context (with call-paths) against an index using the built dist.
+// Usage: node probe-context.mjs <repo-with-.specship> <task words...>
+import { pathToFileURL } from 'node:url';
+import { resolve } from 'node:path';
+
+const [, , repo, ...taskParts] = process.argv;
+const task = taskParts.join(' ');
+if (!repo || !task) { console.error('usage: probe-context.mjs <repo> <task...>'); process.exit(1); }
+
+const load = async (rel) => import(pathToFileURL(resolve(rel)).href);
+const idx = await load('dist/index.js');
+const tools = await load('dist/mcp/tools.js');
+const SpecShip = idx.default?.default ?? idx.default ?? idx.SpecShip;
+const ToolHandler = tools.ToolHandler ?? tools.default?.ToolHandler;
+
+const cg = SpecShip.openSync(repo);
+const h = new ToolHandler(cg);
+const res = await h.execute('specship_context', { task });
+console.log(res.content?.[0]?.text ?? '(no text)');
+try { cg.close?.(); } catch {}

package/scripts/agent-eval/probe-explore.mjs

@@ -0,0 +1,40 @@
+#!/usr/bin/env node
+// One-shot probe: run handleExplore against an existing index using the built
+// dist, print the output + a few stats. Lets us verify explore's coverage fix
+// without a full agent run. Usage: node probe-explore.mjs <repo-with-.specship> "<query>"
+import { pathToFileURL } from 'node:url';
+import { resolve } from 'node:path';
+
+const [, , repo, query] = process.argv;
+if (!repo || !query) {
+  console.error('usage: probe-explore.mjs <repo> "<query>"');
+  process.exit(1);
+}
+
+const load = async (rel) => import(pathToFileURL(resolve(rel)).href);
+const idx = await load('dist/index.js');
+const tools = await load('dist/mcp/tools.js');
+
+// esModuleInterop: dynamic import of CJS yields { default: module.exports, ...named }
+const SpecShip = idx.default?.default ?? idx.default ?? idx.SpecShip;
+const ToolHandler = tools.ToolHandler ?? tools.default?.ToolHandler;
+
+if (typeof SpecShip?.openSync !== 'function') {
+  console.error('could not resolve SpecShip.openSync; index keys:', Object.keys(idx), 'default keys:', idx.default && Object.keys(idx.default));
+  process.exit(2);
+}
+if (typeof ToolHandler !== 'function') {
+  console.error('could not resolve ToolHandler; tools keys:', Object.keys(tools));
+  process.exit(2);
+}
+
+const cg = SpecShip.openSync(repo);
+const h = new ToolHandler(cg);
+const res = await h.execute('specship_explore', { query });
+const text = res.content?.[0]?.text ?? '(no text)';
+console.log(text);
+console.error('\n--- PROBE STATS ---');
+console.error('output chars:', text.length);
+console.error('triggerRender body present (-> setState({})):', /triggerRender[\s\S]{0,400}setState\(\{\}\)/.test(text));
+console.error('App.tsx in source section:', /#### .*App\.tsx —/.test(text));
+try { cg.close?.(); } catch {}

package/scripts/agent-eval/probe-node.mjs

@@ -0,0 +1,20 @@
+#!/usr/bin/env node
+// Probe specship_node (with trail) against an index using the built dist.
+// Usage: node probe-node.mjs <repo-with-.specship> <symbol> [code]
+import { pathToFileURL } from 'node:url';
+import { resolve } from 'node:path';
+
+const [, , repo, symbol, code] = process.argv;
+if (!repo || !symbol) { console.error('usage: probe-node.mjs <repo> <symbol> [code]'); process.exit(1); }
+
+const load = async (rel) => import(pathToFileURL(resolve(rel)).href);
+const idx = await load('dist/index.js');
+const tools = await load('dist/mcp/tools.js');
+const SpecShip = idx.default?.default ?? idx.default ?? idx.SpecShip;
+const ToolHandler = tools.ToolHandler ?? tools.default?.ToolHandler;
+
+const cg = SpecShip.openSync(repo);
+const h = new ToolHandler(cg);
+const res = await h.execute('specship_node', { symbol, includeCode: code === 'code' });
+console.log(res.content?.[0]?.text ?? '(no text)');
+try { cg.close?.(); } catch {}

package/scripts/agent-eval/probe-sweep.mjs

@@ -0,0 +1,119 @@
+#!/usr/bin/env node
+// probe-sweep — direct MCP test across N repos × N tools, no claude needed.
+//
+// Measures response characteristics (size, sections present, signals fired)
+// for each (repo, query) pair against the built dist/. Sub-second per probe;
+// the full sweep below runs in ~10-30s vs hours for a real claude audit.
+//
+// Use this to iterate on backend changes rapidly: change tools.ts /
+// context-builder, npm run build, re-run probe-sweep, compare. Once a
+// change looks good on probe metrics, run a focused claude audit for the
+// few repos that matter to confirm end-to-end cost behavior.
+//
+// Usage: node scripts/agent-eval/probe-sweep.mjs [--tool=context|explore|trace] [--repos=a,b,c]
+import { pathToFileURL } from 'node:url';
+import { resolve } from 'node:path';
+
+const args = Object.fromEntries(
+  process.argv.slice(2).map(a => a.startsWith('--') ? a.slice(2).split('=') : [a, true])
+);
+const TOOL = args.tool ?? 'context';
+
+const load = (rel) => import(pathToFileURL(resolve(rel)).href);
+const idx = await load('dist/index.js');
+const tools = await load('dist/mcp/tools.js');
+const SpecShip = idx.default?.default ?? idx.default ?? idx.SpecShip;
+const ToolHandler = tools.ToolHandler ?? tools.default?.ToolHandler;
+
+// Each entry: repo, query, optional 2nd arg for trace (from, to).
+// The query is the same prompt used in the real claude audits, so probe
+// output is directly comparable to the agent's would-be input.
+const SWEEP = [
+  // Small realworld template repos (the loss cases from the cross-language sweep)
+  { id: 'gin-rw',        repo: '/tmp/specship-corpus/gin-realworld',         q: 'How does this Gin app route a request through its middleware chain to a handler?' },
+  { id: 'go-mux',        repo: '/tmp/specship-corpus/go-mux',                q: 'How does this gorilla/mux app route a request to its handler?' },
+  { id: 'fastapi-rw',    repo: '/tmp/specship-corpus/fastapi-realworld',     q: 'How does FastAPI route a request through its dependencies to a handler?' },
+  { id: 'spring-pc',     repo: '/tmp/specship-corpus/spring-petclinic',      q: 'How does Spring route an HTTP request to a controller method?' },
+  { id: 'axum-rw',       repo: '/tmp/specship-corpus/rust-axum-realworld',   q: 'How does Axum route a request to its handler in this app?' },
+  { id: 'express-rw',    repo: '/tmp/specship-corpus/express-realworld',     q: 'How does this Express app route a request through middleware to a handler?' },
+  { id: 'kotlin-pc',     repo: '/tmp/specship-corpus/kotlin-petclinic',      q: 'How does the Kotlin Spring app route an HTTP request to its handler?' },
+  { id: 'flask-mb',      repo: '/tmp/specship-corpus/flask-microblog',       q: 'How does this Flask app route a request to a view function?' },
+  { id: 'vapor-tpl',     repo: '/tmp/specship-corpus/vapor-template',        q: 'How does Vapor route an HTTP request to its handler?' },
+  { id: 'cpp-leveldb',   repo: '/tmp/specship-corpus/cpp-leveldb',           q: 'How does LevelDB handle a Put operation through to disk?' },
+  { id: 'lualine',       repo: '/tmp/specship-corpus/lualine.nvim',          q: 'How does lualine assemble and render the statusline?' },
+  { id: 'drupal-admin',  repo: '/tmp/specship-corpus/drupal-admintoolbar',   q: 'How does the Drupal admin toolbar module render its toolbar?' },
+  { id: 'svelte-rw',     repo: '/tmp/specship-corpus/svelte-realworld',      q: 'How does this SvelteKit app route a request to a handler?' },
+  { id: 'react-rw',      repo: '/tmp/specship-corpus/react-realworld',       q: 'How does this React app fetch and display articles?' },
+  { id: 'rails-rw',      repo: '/tmp/specship-corpus/rails-realworld',       q: 'How does Rails route a request to a controller action?' },
+  { id: 'flask-rest',    repo: '/tmp/specship-corpus/flask-restful-realworld', q: 'How does Flask-RESTful route a request to a resource method?' },
+  { id: 'laravel-rw',    repo: '/tmp/specship-corpus/laravel-realworld',     q: 'How does Laravel route a request to the controller method?' },
+  { id: 'aspnet-rw',     repo: '/tmp/specship-corpus/aspnet-realworld',      q: 'How does ASP.NET route a request to the controller action?' },
+  // The iter7 wins/ties (to make sure we don't regress)
+  { id: 'cobra',         repo: '/tmp/specship-corpus/cobra',                 q: 'How does cobra parse commands and flags?' },
+  { id: 'sinatra',       repo: '/tmp/specship-corpus/sinatra',               q: 'How does sinatra route a request to its handler?' },
+  { id: 'slim',          repo: '/tmp/specship-corpus/slim',                  q: 'How does slim route a request and apply middleware?' },
+];
+
+// Detect signals in response text — these are the levers we've added that
+// otherwise only show up via "agent ran X more tool calls" downstream.
+const detect = (text) => ({
+  hasEntryPoints: /^### Entry Points/m.test(text),
+  hasRelatedSymbols: /^### Related Symbols/m.test(text),
+  hasFlowTrace: /^## Inline flow trace/m.test(text),
+  hasRouteManifest: /^## Routing manifest/m.test(text),
+  hasTopHandler: /^### Top handler file/m.test(text),
+  hasSmallRepoTail: /This project is small/.test(text),
+});
+
+const filterRepos = args.repos ? new Set(String(args.repos).split(',')) : null;
+const subjects = SWEEP.filter(s => !filterRepos || filterRepos.has(s.id));
+
+const t0 = Date.now();
+const rows = [];
+for (const s of subjects) {
+  try {
+    const cg = SpecShip.openSync(s.repo);
+    const handler = new ToolHandler(cg);
+    const t1 = Date.now();
+    const res = await handler.execute('specship_' + TOOL,
+      TOOL === 'context' ? { task: s.q } :
+      TOOL === 'explore' ? { query: s.q } : { from: 'main', to: 'main' });
+    const text = res.content?.[0]?.text ?? '';
+    const signals = detect(text);
+    rows.push({
+      id: s.id,
+      ms: Date.now() - t1,
+      chars: text.length,
+      lines: text.split('\n').length,
+      ...signals,
+    });
+    try { cg.close?.(); } catch {}
+  } catch (e) {
+    rows.push({ id: s.id, error: String(e).slice(0, 80) });
+  }
+}
+
+// Pretty-print as a compact table.
+const fmt = (r) =>
+  r.error
+    ? `  ${r.id.padEnd(13)} ERROR: ${r.error}`
+    : `  ${r.id.padEnd(13)} ${String(r.chars).padStart(6)}c ${String(r.lines).padStart(4)}L ${String(r.ms).padStart(4)}ms` +
+      ` ${r.hasEntryPoints ? 'EP ' : '   '}` +
+      `${r.hasFlowTrace ? 'TRC ' : '    '}` +
+      `${r.hasRouteManifest ? 'MAN ' : '    '}` +
+      `${r.hasTopHandler ? 'HND ' : '    '}` +
+      `${r.hasSmallRepoTail ? 'TAIL' : '    '}`;
+console.log(`=== probe-sweep tool=${TOOL} n=${subjects.length} (${Date.now() - t0}ms total) ===`);
+console.log('  id            chars  lines    ms signals');
+console.log('  ' + '-'.repeat(56));
+for (const r of rows) console.log(fmt(r));
+
+// Sum + medians for the size pillar
+const sizes = rows.filter(r => !r.error).map(r => r.chars);
+sizes.sort((a, b) => a - b);
+const median = sizes[Math.floor(sizes.length / 2)];
+const sum = sizes.reduce((a, b) => a + b, 0);
+console.log(`  ${'-'.repeat(64)}`);
+console.log(`  median=${median}c  total=${sum}c  ` +
+  `manifest=${rows.filter(r => r.hasRouteManifest).length}/${rows.filter(r => !r.error).length}  ` +
+  `top-handler=${rows.filter(r => r.hasTopHandler).length}/${rows.filter(r => !r.error).length}`);

package/scripts/agent-eval/probe-trace.mjs

@@ -0,0 +1,20 @@
+#!/usr/bin/env node
+// Probe specship_trace against an index using the built dist.
+// Usage: node probe-trace.mjs <repo-with-.specship> <from> <to>
+import { pathToFileURL } from 'node:url';
+import { resolve } from 'node:path';
+
+const [, , repo, from, to] = process.argv;
+if (!repo || !from || !to) { console.error('usage: probe-trace.mjs <repo> <from> <to>'); process.exit(1); }
+
+const load = async (rel) => import(pathToFileURL(resolve(rel)).href);
+const idx = await load('dist/index.js');
+const tools = await load('dist/mcp/tools.js');
+const SpecShip = idx.default?.default ?? idx.default ?? idx.SpecShip;
+const ToolHandler = tools.ToolHandler ?? tools.default?.ToolHandler;
+
+const cg = SpecShip.openSync(repo);
+const h = new ToolHandler(cg);
+const res = await h.execute('specship_trace', { from, to });
+console.log(res.content?.[0]?.text ?? '(no text)');
+try { cg.close?.(); } catch {}

package/scripts/agent-eval/run-agent.sh

@@ -0,0 +1,34 @@
+#!/usr/bin/env bash
+# Headless Claude Code run against a repo with specship MCP, capturing the
+# full stream-json so we can see tool calls + token usage. Complements the
+# interactive itrun.sh: headless gives a clean per-tool breakdown + exact
+# tokens/cost, but defaults to the general-purpose subagent (not Explore).
+# To force the Explore path, ask for it in the prompt.
+#
+# Usage: run-agent.sh <repo-path> <label> "<prompt>"
+# Env: AGENT_EVAL_OUT (default /tmp/agent-eval), CG_BIN (specship dist binary)
+set -uo pipefail
+
+REPO="$1"; LABEL="$2"; PROMPT="$3"
+CG_BIN="${CG_BIN:-$(command -v specship || echo /usr/local/bin/specship)}"
+OUT_DIR="${AGENT_EVAL_OUT:-/tmp/agent-eval}"; mkdir -p "$OUT_DIR"
+OUT="$OUT_DIR/run-${LABEL}.jsonl"
+
+MCP_CONFIG=$(cat <<JSON
+{"mcpServers":{"specship":{"command":"${CG_BIN}","args":["serve","--mcp","--path","${REPO}"]}}}
+JSON
+)
+
+echo "→ running [$LABEL] in $REPO"
+cd "$REPO" || exit 1
+
+claude -p "$PROMPT" \
+  --output-format stream-json --verbose \
+  --permission-mode bypassPermissions \
+  --model opus \
+  --max-budget-usd 2 \
+  --strict-mcp-config --mcp-config "$MCP_CONFIG" \
+  > "$OUT" 2>"$OUT_DIR/run-${LABEL}.err"
+
+echo "exit: $? | wrote $OUT ($(wc -l < "$OUT") lines)"
+node "$(cd "$(dirname "$0")" && pwd)/parse-run.mjs" "$OUT" 2>/dev/null || true