@try-works/dsh-memory 0.1.19 → 0.1.20

package/src/digest.js

@@ -1,255 +1,255 @@
-// Deterministic session -> digest projection, junk classification, and the
-// relevance filter that decides what is even allowed to reach the summarizer.
-// Tolerates the real stored event shapes (data.content blocks, data.message.{role,content}).
-import { clip, gist } from './summarize.js'
-
-/**
- * Extract the assistant answer, not its private reasoning. Reasoning explains how to work; the
- * answer explains what was learned. Research sessions often put the whole report in a later text
- * block after a short reasoning preamble, so concatenating both polluted the candidate and
- * diluted the findings.
- */
-export function assistantAnswerText(content) {
-  if (!Array.isArray(content)) return contentText(content)
-  const answers = content.filter((block) => block && block.type === 'text').map((block) => contentText(block)).filter(Boolean)
-  if (answers.length) return answers.join('\n')
-  // No answer block at all: a message that is still thinking is better than nothing, because a
-  // turn cut short mid-work carries the intent even when it never stated a conclusion.
-  return content.map((block) => contentText(block)).filter(Boolean).join('\n')
-}
-
-/** Recursively pull readable text out of a string, a content block, or an array of blocks. */
-export function contentText(content) {
-  if (typeof content === 'string') return content
-  if (Array.isArray(content)) return content.map(contentText).join('\n')
-  if (content && typeof content === 'object') {
-    if (typeof content.text === 'string') return content.text
-    if (typeof content.content === 'string') return content.content
-  }
-  return ''
-}
-
-/** A path-ish token: at least one separator and a file extension. */
-const PATH_RE = /(?:[A-Za-z]:[\\/])?[A-Za-z0-9_.-]+(?:[\\/][A-Za-z0-9_.-]+)+\.[A-Za-z0-9]{1,6}/g
-
-/**
- * Pull file paths out of arbitrary tool arguments, in order, deduplicated and bounded.
- * Deliberately reads tool CALL arguments only - results are full of paths mentioned in file
- * contents, which say nothing about what the agent chose to work on.
- */
-export function extractPaths(text, limit = 60) {
-  const out = []
-  const seen = new Set()
-  for (const match of String(text ?? '').matchAll(PATH_RE)) {
-    const raw = match[0].replace(/\\/g, '/')
-    if (raw.length > 160 || seen.has(raw)) continue
-    seen.add(raw)
-    out.push(raw)
-    if (out.length >= limit) break
-  }
-  return out
-}
-
-/** Project a raw event list into a digest. Texts stay in memory as candidates; tool
- * calls are reduced to a tally because tool mechanics are never durable memory. */
-export function projectDigest({ sessionId, createdAt, cwd, title }, events) {
-  const digest = {
-    sessionId: sessionId ?? '',
-    title: title ?? '',
-    createdAt: createdAt ?? '',
-    cwd: cwd ?? '',
-    eventCount: (events ?? []).length,
-    userTexts: [],
-    assistantTexts: [],
-    toolTally: {},
-    filesTouched: [],
-    goals: [],
-    checkpoints: [],
-  }
-  for (const event of (events ?? [])) {
-    const type = event.type ?? event.name ?? ''
-    const data = event.data ?? {}
-    if (type === 'user/message') digest.userTexts.push(contentText(data.content))
-    else if (type === 'assistant/message') digest.assistantTexts.push(assistantAnswerText(data.message?.content ?? data.message))
-    else if (type.startsWith('tool/')) {
-      const name = data.toolName ?? data.name ?? type
-      digest.toolTally[name] = (digest.toolTally[name] ?? 0) + 1
-      // what the agent pointed a tool AT is traversal evidence for the codebase map
-      const args = data.arguments === undefined ? data.args : data.arguments
-      const asText = typeof args === 'string' ? args : (args === undefined ? '' : JSON.stringify(args))
-      for (const path of extractPaths(asText, 20)) {
-        if (!digest.filesTouched.includes(path)) digest.filesTouched.push(path)
-      }
-    }
-    else if (type === 'goal/change') digest.goals.push(data.objective ?? '')
-    else if (type === 'compaction/summary') digest.checkpoints.push(data.summary ?? '')
-    else if (type === 'session/title') digest.title = data.title ?? digest.title
-  }
-  digest.userTexts = digest.userTexts.map((t) => t.trim()).filter(Boolean)
-  digest.assistantTexts = digest.assistantTexts.map((t) => t.trim()).filter(Boolean)
-  digest.goals = digest.goals.filter(Boolean)
-  digest.checkpoints = digest.checkpoints.filter(Boolean)
-  digest.filesTouched = digest.filesTouched.slice(-200)
-  return digest
-}
-
-const SELF_MAINTENANCE = /\b(memory_search|memory_remember|memory_sync|memory_associate|memory_workspaces)\b/
-const ACK_RE = /\b(ok|okay|thanks|thank you|got it|done|ack|cool|great|sounds good|will do|nice|sure)\b/i
-
-/** Stamp a digest as trivial (ping), filler (ack-only/self-maintenance), or content. */
-export function classifyJunk(digest) {
-  const userChars = digest.userTexts.join('\n').length
-  const assistantChars = digest.assistantTexts.join('\n').length
-  const userCount = digest.userTexts.length
-  if (digest.eventCount < 100 && userCount <= 2 && assistantChars < 20 && digest.goals.length === 0 && digest.checkpoints.length === 0) {
-    return 'trivial'
-  }
-  const selfChat = digest.userTexts.length > 0 && digest.userTexts.every((t) => SELF_MAINTENANCE.test(t))
-  const ackOnly = userChars < 400 && digest.userTexts.length > 0 && digest.userTexts.every((t) => ACK_RE.test(t))
-  if (selfChat || ackOnly) return 'filler'
-  return 'content'
-}
-
-/** Total tool invocations recorded in a tally. */
-export function totalToolCalls(tally) {
-  let total = 0
-  for (const value of Object.values(tally ?? {})) total += value
-  return total
-}
-
-/** Calendar day (YYYY-MM-DD) of an epoch-ms or ISO timestamp. */
-export function dateOf(value) {
-  if (value == null || value === '') return ''
-  const ms = typeof value === 'number' ? value : Number(value)
-  const date = Number.isFinite(ms) ? new Date(ms) : new Date(String(value))
-  return Number.isNaN(date.getTime()) ? '' : date.toISOString().slice(0, 10)
-}
-
-/** A user turn worth remembering: long enough to carry signal, not a bare acknowledgement. */
-const ACK_ONLY = /^\s*(ok(ay)?|thanks|thank you|got it|done|ack|cool|great|sounds good|will do|nice|sure|yes|no|yep|nope|continue|go on)[.!]?\s*$/i
-
-//
-// System-injected blocks that arrive as USER turns but are not user intent. They are
-// boilerplate the harness or another plugin prepended to the conversation, and they
-// swamped the summarizer: in one research session they were 82% of the material, so a
-// model asked for durable notes mostly saw its own context and answered NO_DURABLE.
-//
-// The DSH memory context entry is OUR OWN injection coming back around. Capturing it
-// is a feedback loop: the system starts treating its own output as user intent, which
-// both wastes the budget and lets stale memory re-enter the store as if it were new.
-//
-export const INJECTED_TURN_PATTERNS = [
-  /^\s*##\s*DSH memory context/m,          // our own injected memory block (feedback loop)
-  /^\s*<system-reminder>/m,                 // skill catalog / harness reminders
-  /^\s*Current runtime context\./m,         // runtime snapshot preamble
-  /^\s*The approval policy changed/m,       // policy notice
-  /^\s*The file .{0,200}? has been (modified|created|deleted)/m,
-  /^\s*\[Request interrupted by user/m,
-  /^\s*<available_skills>/m,
-]
-
-/**
- * Remove injected blocks that ride inside an otherwise-real turn. A turn can carry a genuine
- * user request AND a prepended context block; dropping the whole turn would lose the request,
- * and keeping the block wastes the budget, so the block is excised and the remainder kept.
- */
-export function stripInjectedBlocks(text) {
-  let value = String(text ?? '')
-  value = value.replace(/^\s*##\s*DSH memory context[\s\S]*?(?=\n\s*\n(?![-*#]|\s)|$)/m, '')
-  value = value.replace(/<system-reminder>[\s\S]*?<\/system-reminder>/g, '')
-  value = value.replace(/<available_skills>[\s\S]*?<\/available_skills>/g, '')
-  value = value.replace(/^\s*Current runtime context\.[\s\S]*?(?=\n\s*\n|$)/m, '')
-  value = value.replace(/^\s*The approval policy changed[^\n]*\n?/m, '')
-  return value.replace(/\n{3,}/g, '\n\n').trim()
-}
-
-/** True when a turn is harness-injected context rather than something the user said. */
-export function isInjectedTurn(text) {
-  const value = String(text ?? '')
-  if (value === '') return false
-  return INJECTED_TURN_PATTERNS.some((re) => re.test(value))
-}
-
-/** Rough byte weight of a candidate set (what the summarizer prompt will carry). */
-export function candidateBytes(candidates) {
-  let total = 0
-  for (const list of [candidates.userTurns, candidates.assistantGists, candidates.goals, candidates.checkpoints, candidates.files]) {
-    for (const item of list ?? []) total += item.length + 3
-  }
-  return total
-}
-
-/** Trim the OLDEST low-signal material until the candidate set fits the byte budget. */
-function boundCandidates(candidates, maxBytes) {
-  let guard = 100000
-  while (candidateBytes(candidates) > maxBytes && guard-- > 0) {
-    if (candidates.assistantGists.length > 0) { candidates.assistantGists.shift(); continue }
-    if (candidates.userTurns.length > 1) { candidates.userTurns.shift(); continue }
-    if (candidates.checkpoints.length > 1) { candidates.checkpoints.shift(); continue }
-    if ((candidates.files ?? []).length > 1) { candidates.files.shift(); continue }
-    if (candidates.goals.length > 1) { candidates.goals.shift(); continue }
-    // Last resort: hard-clip the single remaining item so the budget is a guarantee,
-    // never a hint — an oversized lone turn must not reach the summarizer whole.
-    const over = candidateBytes(candidates) - maxBytes
-    const last = candidates.userTurns[candidates.userTurns.length - 1]
-    if (last !== undefined && last.length > over + 8) {
-      candidates.userTurns[candidates.userTurns.length - 1] = clip(last, last.length - over - 4)
-      continue
-    }
-    break
-  }
-  return candidates
-}
-
-/**
- * The relevance filter: bounded, deterministic material for the summarizer.
- * Relevance is decided in two layers — this filter drops tool mechanics, bare
- * acknowledgements, and everything outside a recent window; the durable summary
- * instructions then decide what is actually durable.
- */
-export function extractCandidates(digest, options = {}) {
-  const opts = {
-    maxUserTurns: 60, maxAssistantGists: 40, maxMessageChars: 600,
-    maxGoalChars: 200, maxCheckpoints: 4, maxFiles: 60, maxBytes: 40000, date: '',
-    ...options,
-  }
-  const candidates = {
-    sessionId: digest.sessionId,
-    title: digest.title,
-    date: opts.date || dateOf(digest.createdAt),
-    cwd: digest.cwd,
-    eventCount: digest.eventCount,
-    toolCallCount: totalToolCalls(digest.toolTally),
-    goals: (digest.goals ?? []).slice(-3).map((g) => clip(g, opts.maxGoalChars)).filter(Boolean),
-    checkpoints: (digest.checkpoints ?? []).slice(-opts.maxCheckpoints).map((c) => clip(c, opts.maxMessageChars)).filter(Boolean),
-    userTurns: (digest.userTexts ?? [])
-      // Injected context is dropped here, not merely down-weighted: it is not user intent, and
-      // leaving it in crowds out the real turns inside a bounded prompt.
-      .filter((t) => t.length > 12 && !ACK_ONLY.test(t) && !isInjectedTurn(t))
-      .map(stripInjectedBlocks)
-      .slice(-opts.maxUserTurns)
-      .map((t) => clip(t, opts.maxMessageChars)),
-    // Traversal evidence: the files this session actually worked in.
-    files: (digest.filesTouched ?? []).slice(-opts.maxFiles).map((f) => clip(f, 120)),
-    assistantGists: (digest.assistantTexts ?? [])
-      .slice(-opts.maxAssistantGists)
-      .map((t) => gist(t, Math.min(240, opts.maxMessageChars)))
-      .filter(Boolean),
-  }
-  return boundCandidates(candidates, opts.maxBytes)
-}
-
-/** Render the stable digest-file header (metadata only — never message bodies). */
-export function renderDigestHeader(meta) {
-  return [
-    '# Session ' + (meta.sessionId || ''),
-    'Title: ' + (meta.title || ''),
-    'Cwd: ' + (meta.cwd || ''),
-    'Created: ' + (meta.createdAt || ''),
-    'Date: ' + (meta.date || ''),
-    'Events: ' + (meta.eventCount == null ? 0 : meta.eventCount),
-    'Tool calls: ' + (meta.toolCallCount == null ? 0 : meta.toolCallCount),
-    'Captured: ' + (meta.capturedAt || ''),
-    '',
-  ].join('\n')
-}
+// Deterministic session -> digest projection, junk classification, and the
+// relevance filter that decides what is even allowed to reach the summarizer.
+// Tolerates the real stored event shapes (data.content blocks, data.message.{role,content}).
+import { clip, gist } from './summarize.js'
+
+/**
+ * Extract the assistant answer, not its private reasoning. Reasoning explains how to work; the
+ * answer explains what was learned. Research sessions often put the whole report in a later text
+ * block after a short reasoning preamble, so concatenating both polluted the candidate and
+ * diluted the findings.
+ */
+export function assistantAnswerText(content) {
+  if (!Array.isArray(content)) return contentText(content)
+  const answers = content.filter((block) => block && block.type === 'text').map((block) => contentText(block)).filter(Boolean)
+  if (answers.length) return answers.join('\n')
+  // No answer block at all: a message that is still thinking is better than nothing, because a
+  // turn cut short mid-work carries the intent even when it never stated a conclusion.
+  return content.map((block) => contentText(block)).filter(Boolean).join('\n')
+}
+
+/** Recursively pull readable text out of a string, a content block, or an array of blocks. */
+export function contentText(content) {
+  if (typeof content === 'string') return content
+  if (Array.isArray(content)) return content.map(contentText).join('\n')
+  if (content && typeof content === 'object') {
+    if (typeof content.text === 'string') return content.text
+    if (typeof content.content === 'string') return content.content
+  }
+  return ''
+}
+
+/** A path-ish token: at least one separator and a file extension. */
+const PATH_RE = /(?:[A-Za-z]:[\\/])?[A-Za-z0-9_.-]+(?:[\\/][A-Za-z0-9_.-]+)+\.[A-Za-z0-9]{1,6}/g
+
+/**
+ * Pull file paths out of arbitrary tool arguments, in order, deduplicated and bounded.
+ * Deliberately reads tool CALL arguments only - results are full of paths mentioned in file
+ * contents, which say nothing about what the agent chose to work on.
+ */
+export function extractPaths(text, limit = 60) {
+  const out = []
+  const seen = new Set()
+  for (const match of String(text ?? '').matchAll(PATH_RE)) {
+    const raw = match[0].replace(/\\/g, '/')
+    if (raw.length > 160 || seen.has(raw)) continue
+    seen.add(raw)
+    out.push(raw)
+    if (out.length >= limit) break
+  }
+  return out
+}
+
+/** Project a raw event list into a digest. Texts stay in memory as candidates; tool
+ * calls are reduced to a tally because tool mechanics are never durable memory. */
+export function projectDigest({ sessionId, createdAt, cwd, title }, events) {
+  const digest = {
+    sessionId: sessionId ?? '',
+    title: title ?? '',
+    createdAt: createdAt ?? '',
+    cwd: cwd ?? '',
+    eventCount: (events ?? []).length,
+    userTexts: [],
+    assistantTexts: [],
+    toolTally: {},
+    filesTouched: [],
+    goals: [],
+    checkpoints: [],
+  }
+  for (const event of (events ?? [])) {
+    const type = event.type ?? event.name ?? ''
+    const data = event.data ?? {}
+    if (type === 'user/message') digest.userTexts.push(contentText(data.content))
+    else if (type === 'assistant/message') digest.assistantTexts.push(assistantAnswerText(data.message?.content ?? data.message))
+    else if (type.startsWith('tool/')) {
+      const name = data.toolName ?? data.name ?? type
+      digest.toolTally[name] = (digest.toolTally[name] ?? 0) + 1
+      // what the agent pointed a tool AT is traversal evidence for the codebase map
+      const args = data.arguments === undefined ? data.args : data.arguments
+      const asText = typeof args === 'string' ? args : (args === undefined ? '' : JSON.stringify(args))
+      for (const path of extractPaths(asText, 20)) {
+        if (!digest.filesTouched.includes(path)) digest.filesTouched.push(path)
+      }
+    }
+    else if (type === 'goal/change') digest.goals.push(data.objective ?? '')
+    else if (type === 'compaction/summary') digest.checkpoints.push(data.summary ?? '')
+    else if (type === 'session/title') digest.title = data.title ?? digest.title
+  }
+  digest.userTexts = digest.userTexts.map((t) => t.trim()).filter(Boolean)
+  digest.assistantTexts = digest.assistantTexts.map((t) => t.trim()).filter(Boolean)
+  digest.goals = digest.goals.filter(Boolean)
+  digest.checkpoints = digest.checkpoints.filter(Boolean)
+  digest.filesTouched = digest.filesTouched.slice(-200)
+  return digest
+}
+
+const SELF_MAINTENANCE = /\b(memory_search|memory_remember|memory_sync|memory_associate|memory_workspaces)\b/
+const ACK_RE = /\b(ok|okay|thanks|thank you|got it|done|ack|cool|great|sounds good|will do|nice|sure)\b/i
+
+/** Stamp a digest as trivial (ping), filler (ack-only/self-maintenance), or content. */
+export function classifyJunk(digest) {
+  const userChars = digest.userTexts.join('\n').length
+  const assistantChars = digest.assistantTexts.join('\n').length
+  const userCount = digest.userTexts.length
+  if (digest.eventCount < 100 && userCount <= 2 && assistantChars < 20 && digest.goals.length === 0 && digest.checkpoints.length === 0) {
+    return 'trivial'
+  }
+  const selfChat = digest.userTexts.length > 0 && digest.userTexts.every((t) => SELF_MAINTENANCE.test(t))
+  const ackOnly = userChars < 400 && digest.userTexts.length > 0 && digest.userTexts.every((t) => ACK_RE.test(t))
+  if (selfChat || ackOnly) return 'filler'
+  return 'content'
+}
+
+/** Total tool invocations recorded in a tally. */
+export function totalToolCalls(tally) {
+  let total = 0
+  for (const value of Object.values(tally ?? {})) total += value
+  return total
+}
+
+/** Calendar day (YYYY-MM-DD) of an epoch-ms or ISO timestamp. */
+export function dateOf(value) {
+  if (value == null || value === '') return ''
+  const ms = typeof value === 'number' ? value : Number(value)
+  const date = Number.isFinite(ms) ? new Date(ms) : new Date(String(value))
+  return Number.isNaN(date.getTime()) ? '' : date.toISOString().slice(0, 10)
+}
+
+/** A user turn worth remembering: long enough to carry signal, not a bare acknowledgement. */
+const ACK_ONLY = /^\s*(ok(ay)?|thanks|thank you|got it|done|ack|cool|great|sounds good|will do|nice|sure|yes|no|yep|nope|continue|go on)[.!]?\s*$/i
+
+//
+// System-injected blocks that arrive as USER turns but are not user intent. They are
+// boilerplate the harness or another plugin prepended to the conversation, and they
+// swamped the summarizer: in one research session they were 82% of the material, so a
+// model asked for durable notes mostly saw its own context and answered NO_DURABLE.
+//
+// The DSH memory context entry is OUR OWN injection coming back around. Capturing it
+// is a feedback loop: the system starts treating its own output as user intent, which
+// both wastes the budget and lets stale memory re-enter the store as if it were new.
+//
+export const INJECTED_TURN_PATTERNS = [
+  /^\s*##\s*DSH memory context/m,          // our own injected memory block (feedback loop)
+  /^\s*<system-reminder>/m,                 // skill catalog / harness reminders
+  /^\s*Current runtime context\./m,         // runtime snapshot preamble
+  /^\s*The approval policy changed/m,       // policy notice
+  /^\s*The file .{0,200}? has been (modified|created|deleted)/m,
+  /^\s*\[Request interrupted by user/m,
+  /^\s*<available_skills>/m,
+]
+
+/**
+ * Remove injected blocks that ride inside an otherwise-real turn. A turn can carry a genuine
+ * user request AND a prepended context block; dropping the whole turn would lose the request,
+ * and keeping the block wastes the budget, so the block is excised and the remainder kept.
+ */
+export function stripInjectedBlocks(text) {
+  let value = String(text ?? '')
+  value = value.replace(/^\s*##\s*DSH memory context[\s\S]*?(?=\n\s*\n(?![-*#]|\s)|$)/m, '')
+  value = value.replace(/<system-reminder>[\s\S]*?<\/system-reminder>/g, '')
+  value = value.replace(/<available_skills>[\s\S]*?<\/available_skills>/g, '')
+  value = value.replace(/^\s*Current runtime context\.[\s\S]*?(?=\n\s*\n|$)/m, '')
+  value = value.replace(/^\s*The approval policy changed[^\n]*\n?/m, '')
+  return value.replace(/\n{3,}/g, '\n\n').trim()
+}
+
+/** True when a turn is harness-injected context rather than something the user said. */
+export function isInjectedTurn(text) {
+  const value = String(text ?? '')
+  if (value === '') return false
+  return INJECTED_TURN_PATTERNS.some((re) => re.test(value))
+}
+
+/** Rough byte weight of a candidate set (what the summarizer prompt will carry). */
+export function candidateBytes(candidates) {
+  let total = 0
+  for (const list of [candidates.userTurns, candidates.assistantGists, candidates.goals, candidates.checkpoints, candidates.files]) {
+    for (const item of list ?? []) total += item.length + 3
+  }
+  return total
+}
+
+/** Trim the OLDEST low-signal material until the candidate set fits the byte budget. */
+function boundCandidates(candidates, maxBytes) {
+  let guard = 100000
+  while (candidateBytes(candidates) > maxBytes && guard-- > 0) {
+    if (candidates.assistantGists.length > 0) { candidates.assistantGists.shift(); continue }
+    if (candidates.userTurns.length > 1) { candidates.userTurns.shift(); continue }
+    if (candidates.checkpoints.length > 1) { candidates.checkpoints.shift(); continue }
+    if ((candidates.files ?? []).length > 1) { candidates.files.shift(); continue }
+    if (candidates.goals.length > 1) { candidates.goals.shift(); continue }
+    // Last resort: hard-clip the single remaining item so the budget is a guarantee,
+    // never a hint — an oversized lone turn must not reach the summarizer whole.
+    const over = candidateBytes(candidates) - maxBytes
+    const last = candidates.userTurns[candidates.userTurns.length - 1]
+    if (last !== undefined && last.length > over + 8) {
+      candidates.userTurns[candidates.userTurns.length - 1] = clip(last, last.length - over - 4)
+      continue
+    }
+    break
+  }
+  return candidates
+}
+
+/**
+ * The relevance filter: bounded, deterministic material for the summarizer.
+ * Relevance is decided in two layers — this filter drops tool mechanics, bare
+ * acknowledgements, and everything outside a recent window; the durable summary
+ * instructions then decide what is actually durable.
+ */
+export function extractCandidates(digest, options = {}) {
+  const opts = {
+    maxUserTurns: 60, maxAssistantGists: 40, maxMessageChars: 600,
+    maxGoalChars: 200, maxCheckpoints: 4, maxFiles: 60, maxBytes: 40000, date: '',
+    ...options,
+  }
+  const candidates = {
+    sessionId: digest.sessionId,
+    title: digest.title,
+    date: opts.date || dateOf(digest.createdAt),
+    cwd: digest.cwd,
+    eventCount: digest.eventCount,
+    toolCallCount: totalToolCalls(digest.toolTally),
+    goals: (digest.goals ?? []).slice(-3).map((g) => clip(g, opts.maxGoalChars)).filter(Boolean),
+    checkpoints: (digest.checkpoints ?? []).slice(-opts.maxCheckpoints).map((c) => clip(c, opts.maxMessageChars)).filter(Boolean),
+    userTurns: (digest.userTexts ?? [])
+      // Injected context is dropped here, not merely down-weighted: it is not user intent, and
+      // leaving it in crowds out the real turns inside a bounded prompt.
+      .filter((t) => t.length > 12 && !ACK_ONLY.test(t) && !isInjectedTurn(t))
+      .map(stripInjectedBlocks)
+      .slice(-opts.maxUserTurns)
+      .map((t) => clip(t, opts.maxMessageChars)),
+    // Traversal evidence: the files this session actually worked in.
+    files: (digest.filesTouched ?? []).slice(-opts.maxFiles).map((f) => clip(f, 120)),
+    assistantGists: (digest.assistantTexts ?? [])
+      .slice(-opts.maxAssistantGists)
+      .map((t) => gist(t, Math.min(240, opts.maxMessageChars)))
+      .filter(Boolean),
+  }
+  return boundCandidates(candidates, opts.maxBytes)
+}
+
+/** Render the stable digest-file header (metadata only — never message bodies). */
+export function renderDigestHeader(meta) {
+  return [
+    '# Session ' + (meta.sessionId || ''),
+    'Title: ' + (meta.title || ''),
+    'Cwd: ' + (meta.cwd || ''),
+    'Created: ' + (meta.createdAt || ''),
+    'Date: ' + (meta.date || ''),
+    'Events: ' + (meta.eventCount == null ? 0 : meta.eventCount),
+    'Tool calls: ' + (meta.toolCallCount == null ? 0 : meta.toolCallCount),
+    'Captured: ' + (meta.capturedAt || ''),
+    '',
+  ].join('\n')
+}