job-application-agent 3.7.0 → 3.7.2

package/job-application-agent/SKILL.md

@@ -37,12 +37,16 @@ copyable text to the candidate for manual sending, and record actual observation
 Do not automate LinkedIn/X access or messaging. Existing application autonomy
 does not enable this module. Outreach commands bypass analytics and community
 transmissions; do not run updater/telemetry/community commands as part of an
-outreach-only workflow. Keep outreach separate from applications and never
-classify a proposed screen as scheduled. No scheduled follow-up is created.
+outreach-only workflow. Keep drafting and sending separate from ATS forms.
+A `sent-verified` outreach counts toward the active application round the same
+way a confirmed apply does, unless that company is already a confirmed apply on
+the same round. Clear does not un-confirm. A later `not-sent` or `failed`
+delivery correction does. Never classify a proposed screen as scheduled. No scheduled
+follow-up is created.
 
 ## Accounting
 
-Read [references/ACCOUNTING.md](references/ACCOUNTING.md) before recording delivery evidence, recovery attempts, or per-lead discovery. For new rounds, record each lead with `round lead --stdin` and derive source totals from those records. Email access is optional: visible browser success counts, verified email sends count with receipt unknown, and matched final delivery failures correct effective totals. Preserve historical events and use explicit corrections for conflicts.
+Read [references/ACCOUNTING.md](references/ACCOUNTING.md) before recording delivery evidence, recovery attempts, or per-lead discovery. For new rounds, record each lead with `round lead --stdin` and derive source totals from those records. Email access is optional: visible browser success counts, verified email sends count with receipt unknown, and matched final delivery failures correct effective totals. A sent-verified outreach counts toward the same round `confirmedCount` unless that company already has a counted apply. Preserve historical events and use explicit corrections for conflicts.
 
 ## Discover and assess
 
@@ -76,7 +80,7 @@ Check `round status` after the initial discovery pass and before submitting. Pre
 
 1. When private cloud state is configured, run `cloud status`, acquire the application-run lease with `cloud lease-acquire`, and renew it at least every five minutes. A client without the live lease may research and draft but must not submit.
 2. Recheck employer, title, direct domain, posting status, eligibility, and `autoEligible` immediately before submission.
-3. Run `ledger check --stdin` with the internal ledger ID, canonical URL, employer job ID, company, and role when available. Review both requisition duplicate status and same-company history.
+3. Run `ledger check --stdin` with any one identifier set: job URL, internal application id, employer job id plus company, or company+role. Include more identifiers when known. A company+role match is a possible duplicate and returns the stored URL; never treat it as a hard already-applied. Review both requisition duplicate status and same-company history.
 4. Stop on a hard ledger-ID, canonical-URL, employer-job-ID, or requisition duplicate. Treat a same-company/same-role alias as a possible duplicate. Use `duplicateOverride: "NEW REQUISITION CONFIRMED"` only after verifying it is a distinct requisition.
 5. For a genuinely different role at a previously applied company, follow `companyReapply`: proceed automatically only when it returns `eligible-after-cooldown` (15 full days since the latest company application and no recorded outcome). `cooldown-active` and `follow-up-present` require the candidate's explicit approval and `companyReapplyOverride: "CANDIDATE APPROVED EARLY REAPPLICATION"`.
 6. Keep authentication in the existing browser session. Never inspect cookies, local storage, passwords, or session files.
@@ -95,6 +99,7 @@ Check `round status` after the initial discovery pass and before submitting. Pre
 
 - Keep `applications.ndjson` and `outcomes.ndjson` append-only. Never delete or rewrite historical rows.
 - Record outcomes with `ledger outcome --stdin`. Use structured rejection reasons and mark each as `explicit` or `inferred`. Do not treat an inference as a candidate fact.
+- Run `ledger check` and `ledger outcome` as two separate CLI processes. Do not combine them in one invocation. When mail has company and role but no URL or id, look up the row with `ledger check` first; that hit is only a possible duplicate and includes the stored URL and id. Then pass the returned `match.id` to `ledger outcome`. If `match.id` is absent, stop and ask; do not guess among fuzzy company+role hits.
 - After an interview, optionally record `interviewQuality` (`promising`, `viable`, `weak`, or `dead`) and a bounded `failurePoint`. Keep free-form interview notes private.
 - Rely on idempotent outcome recording; identical events do not append rows or emit duplicate telemetry.
 - Audit matched delivery failures with authorized email tools when available; otherwise report delivery not audited and continue. Keep delivery failures separate from hiring rejections.
@@ -131,7 +136,7 @@ node scripts/job-application.mjs ledger review
 node scripts/job-application.mjs ledger review-ack --stdin
 node scripts/job-application.mjs autonomy grant --stdin
 node scripts/job-application.mjs autonomy status|preview|revoke
-node scripts/job-application.mjs round start|source|complete --stdin
+node scripts/job-application.mjs round start|source|confirm|complete --stdin
 node scripts/job-application.mjs round status [round-id]
 node scripts/job-application.mjs sources list [--stdin]
 node scripts/job-application.mjs sources jobs [--stdin]

package/job-application-agent/references/ACCOUNTING.md

@@ -6,7 +6,7 @@ Browser applications count after visible ATS success. They do not require email
 
 Match a final delivery failure to the actual application attempt using its sent message, recipient, timing and returned failure evidence. Temporary delays, unrelated bounces, and hiring rejections are not delivery failures. A failed notification cannot overturn independent browser confirmation. Never guess an alternate address or automatically resend an uncertain transmission.
 
-`ledger review`, `round status`, and `round complete` share one effective-count projection. They expose `recordedSubmissionCount`, `effectiveSubmissionCount`, `failedDeliveryCount`, and `receiptUnknownEmailCount`. Existing `submittedTotal`/`confirmedCount` fields now reflect effective submissions, not proven email receipt. Conversion denominators exclude failed applications while historical outcomes remain visible. Review cadence and acknowledgement checkpoints use recorded canonical submissions and recorded mature applications, so a late failure does not postpone the next review. Public usage events remain historical activity metrics.
+`ledger review`, `round status`, and `round complete` share one effective-count projection for ledger applications. They expose `recordedSubmissionCount`, `effectiveSubmissionCount`, `failedDeliveryCount`, and `receiptUnknownEmailCount`. Existing `submittedTotal`/`confirmedCount` fields now reflect effective submissions, not proven email receipt. `round status` also adds sent-verified outreach that is attached to the round and not already covered by a counted apply at the same company, exposed as `applyConfirmationCount` plus `outreachConfirmationCount`. Clearing outreach content never drops that attached count. An explicit delivery correction to `not-sent` or `failed` does: the durable confirmation stays on disk, but `confirmedCount` stops counting that send. Conversion denominators exclude failed applications while historical outcomes remain visible. Review cadence and acknowledgement checkpoints use recorded canonical submissions and recorded mature applications, so a late failure does not postpone the next review. Public usage events remain historical activity metrics.
 
 A late failure preserves the original completion event. `shortfallCount` and `needsRecovery` flag the deficit; no run starts or round reopens automatically. Review the derived attention queue before a later authorized recovery.
 

package/job-application-agent/references/ANALYTICS.md

@@ -65,7 +65,7 @@ Company and title values are bounded and rejected when they resemble an email, p
 | `application_paused` | Job hash, ATS, stage, bounded reason |
 | `application_skipped` | Job hash, bounded reason, fit score, eligibility |
 | `application_submitted` | Company, title, job hash/domain, ATS, duration, fields filled, short-answer count, resume-upload Boolean, approval mode |
-| `round_completed` | Requested/submitted/assessed/skipped/paused/error counts, duration bucket; optional attempted/searched/blocked source counts, maximum source share percentage, bounded concentration reason |
+| `round_completed` | Requested/submitted/assessed/skipped/paused/error counts, duration bucket; optional attempted/searched/blocked source counts, maximum source share percentage, bounded concentration reason. `submittedCount` and `assessedCount` are ledger application totals only and never include attached outreach. |
 | `outcome_recorded` | Company, title, job hash/domain, ATS, outcome, days since submission, optional bounded interview quality/failure point |
 | `review_generated` | Canonical unique-submission and outcome counts, review-due Boolean |
 | `skill_error` | Stable error code, workflow stage, ATS/job hash when available, recoverable Boolean |

package/job-application-agent/references/OUTREACH.md

@@ -173,8 +173,15 @@ are explicitly stale. On reconnect, the cache is replaced by authoritative cloud
 data, including clear tombstones. If a lower server revision suggests a restored
 backend or delayed response, the cache is discarded; another online read is
 required before offline inspection is available. Handoffs use atomic database reservations,
-not application submission leases. No operation changes application counts,
-round progress, or the existing application-autonomy grant.
+not application submission leases. Recording `sent-verified` increments the
+active application round (or the round already attached to that opportunity)
+when that company does not already have a counted apply on the same round.
+Auto-attach first does a targeted authenticated reconcile of cloud `rounds`
+events so a second or stale host can see the open round. Apply plus outreach
+at the same company is one confirmation, not two. The existing
+application-autonomy grant is unchanged. Use `round confirm --stdin`
+with `{ "roundId":"round-...","outreachId":"opportunity-..." }` to attach an
+already sent-verified opportunity to an open round.
 
 Outreach commands bypass telemetry, identity transmission, community sharing,
 and their automatic retry paths—including command errors. Only authenticated
@@ -183,11 +190,17 @@ evidence stay in sensitive content storage. Audit rows retain opaque IDs,
 timestamps, coarse transitions, and private keyed fingerprints.
 
 Clear input: `{"operationId":"clear-1","ids":["opportunity-1"]}`.
-Clearing removes sensitive content and blocks resurrection. Minimal outcomes,
-audit history, company/contact fingerprints and suppression remain. Do not
-describe this as deleting every personal-data trace. Disconnected host caches
-and user-managed exports cannot be remotely erased; backups retain old content
-until their existing 30-day expiry.
+Clearing removes sensitive content and blocks resurrection. It never
+un-confirms a counted round attachment: `confirmedCount` stays on the durable
+`submission-confirmed` event and uses the company already stored there.
+A later `not-sent` or `failed` correction of that send is different: the
+attachment event remains, but the projection drops it from `confirmedCount`.
+New attachments require a live outreach snapshot when cloud state is
+configured; cached reads may still power `round status`. Minimal
+outcomes, audit history, company/contact fingerprints and suppression remain.
+Do not describe this as deleting every personal-data trace. Disconnected host
+caches and user-managed exports cannot be remotely erased; backups retain old
+content until their existing 30-day expiry.
 
 Private backups include outreach tables. `cloud export` also includes the
 outreach snapshot and deletion manifest. A manifest must be obtained from the

package/job-application-agent/references/RUNS.md

@@ -14,6 +14,7 @@ Start each batch with an explicit ID:
 node scripts/job-application.mjs round start --stdin
 node scripts/job-application.mjs round source --stdin
 node scripts/job-application.mjs round status [round-id]
+node scripts/job-application.mjs round confirm --stdin
 node scripts/job-application.mjs round complete --stdin
 ```
 
@@ -62,7 +63,7 @@ Coverage reports emit bounded `source_checked` analytics automatically. Only all
 Before completion, audit final email delivery failures through authorized email tools when available; otherwise report delivery not audited and continue. Verified email sends count with receipt unknown. Use `ledger delivery` to record matched failures, never `ledger outcome rejected`. Effective totals exclude failed attempts; completed rounds retain their completion record and expose a recovery shortfall. See [ACCOUNTING.md](ACCOUNTING.md) for safe linked replacement attempts.
 
 
-Count only unique applications with a visible employer/ATS confirmation or a verified sent recruiting email that were also added to the ledger with the same `roundId`. Filled forms, blockers, drafts, unsent email, and ambiguous confirmations never count. `round complete` rejects an under-target round.
+Count unique applications with a visible employer/ATS confirmation or a verified sent recruiting email that were also added to the ledger with the same `roundId`, plus sent-verified outreach attached to that round when the company is not already a counted apply. Filled forms, blockers, drafts, unsent email, user-reported-only outreach, and ambiguous confirmations never count. Apply plus outreach at the same company is one confirmation. `round complete` rejects an under-target round.
 
 Run both company-level and requisition-level duplicate checks before filling and again immediately before transmission. Hard ledger-ID, canonical-URL, employer-job-ID, and requisition duplicates always stop. Same-role aliases require a verified distinct requisition and `NEW REQUISITION CONFIRMED`. A genuinely different role at the same company may proceed automatically only when `companyReapply.decision` is `eligible-after-cooldown`: 15 full days have passed since the latest company application and no outcome has been recorded. `cooldown-active` and `follow-up-present` require explicit candidate approval.
 

package/job-application-agent/references/SCHEMAS.md

@@ -81,7 +81,7 @@ Assessment output retains `review`, `ask`, `skip`, and `exclude`, and adds `auto
 
 ## Duplicate check input
 
-Include as many identifiers as are known.
+`ledger check` accepts any one identifier set: job URL, internal application id, employer job id plus company, or company+role. Include more identifiers when known. Company-only, role-only, and empty objects are rejected.
 
 ```json
 {
@@ -93,7 +93,7 @@ Include as many identifiers as are known.
 }
 ```
 
-The check removes fragments and non-job query parameters while retaining recognized job or requisition identifiers. Matching ledger ID, canonical URL, or same-company employer job ID is a hard duplicate. Same company and role without a shared job ID is a possible duplicate.
+The check removes fragments and non-job query parameters while retaining recognized job or requisition identifiers. Matching ledger ID, canonical URL, or same-company employer job ID is a hard duplicate. Same company and role without a shared job ID is a possible duplicate and includes the stored URL. Never promote a company+role match to a hard already-applied. `ledger add` still requires a real job URL.
 
 `ledger check` also returns `companyReapply`. A genuinely different role is `eligible-after-cooldown` only when at least 15 full days have passed since the latest application to that company and no outcome has been recorded for that application. Hard duplicates are never eligible. Same-role matches, `cooldown-active`, and `follow-up-present` remain blocked at `ledger add` unless their exact documented override is present.
 
@@ -155,7 +155,7 @@ Use `round source --stdin` for per-source search/blocker reports and optional at
 { "requestedCount": 30 }
 ```
 
-`round start --stdin` appends a `started` event to owner-only `rounds.ndjson` and returns a generated `roundId`. Add that ID to every confirmed ledger entry. `round complete --stdin` accepts `{ "roundId": "round-..." }`, plus concentration reason and evidence when required, and appends a completion event only after the target count, source coverage, attribution, and concentration requirements are satisfied.
+`round start --stdin` appends a `started` event to owner-only `rounds.ndjson` and returns a generated `roundId`. Add that ID to every confirmed ledger entry. A `sent-verified` outreach also counts toward `confirmedCount` for the active or attached open round when that company is not already a counted apply on the same round. Once attached, that confirmation stays after `outreach clear`; clear wipes message text and PII only. A later delivery correction to `not-sent` or `failed` stops counting that send. `round confirm` and auto-attach require an authoritative outreach snapshot; they do not mutate from a stale offline cache. `round confirm --stdin` accepts `{ "roundId": "round-...", "outreachId": "opportunity-..." }` (or `outreachIds`) to attach already sent-verified outreach without creating a second ledger application. `round complete --stdin` accepts `{ "roundId": "round-..." }`, plus concentration reason and evidence when required, and appends a completion event only after the target count, source coverage, attribution, and concentration requirements are satisfied. Discovery attribution and lead linkage still apply only to ledger applications.
 
 ## Attention input
 

package/job-application-agent/scripts/cloud-state-client.mjs

@@ -338,10 +338,17 @@ export class CloudStateClient {
     return (await this.request(`/v2/intents/${encodeURIComponent(intentId)}/confirm`, { method: 'POST', headers: { 'content-type': 'application/json' }, body: JSON.stringify({ application, leaseId, idempotencyKey }) })).json();
   }
 
-  async reconcile({ dryRun = true, provenance = 'local-reconcile' } = {}) {
+  async reconcile({ dryRun = true, provenance = 'local-reconcile', streams = null } = {}) {
     await this.requireAccounting();
     const report = { dryRun, streams: {}, imported: 0, downloaded: 0 };
-    for (const [stream, filename] of Object.entries(CLOUD_STREAM_FILES)) {
+    const selected = streams == null
+      ? Object.entries(CLOUD_STREAM_FILES)
+      : streams.map((stream) => {
+          const filename = CLOUD_STREAM_FILES[stream];
+          if (!filename) throw new Error(`Unknown cloud stream: ${stream}.`);
+          return [stream, filename];
+        });
+    for (const [stream, filename] of selected) {
       if (!dryRun) await this.flushAccountingWrites(stream);
       const normalize = value => stream === 'delivery' ? validateDelivery(value) : stream === 'discovery' && value.version === 1 && value.type === 'lead-reviewed' ? validateLead(value) : value;
       let local = (await readNdjson(join(this.stateDir, filename))).map(normalize);

package/job-application-agent/scripts/job-application.mjs

@@ -26,6 +26,7 @@ import {
   extractNarrativeQuestionsFromText,
   normalizeAttentionQuestions,
 } from './attention-questions.mjs';
+import { confirmOutreachTowardRound, loadOutreachRoundView, loadSentVerifiedOutreach, outreachConfirmationEvents, projectOutreachRoundCounts, resolveOutreachRoundId } from './outreach-round.mjs';
 
 const SOURCES = new Set(['linkedin', 'greenhouse', 'lever', 'ashby', 'workable', 'comeet', 'workday', 'rippling', 'smartrecruiters', 'google-form', 'company', 'email', 'other']);
 const DISCOVERY_SOURCES = new Set(['direct-company', 'linkedin', 'x', 'yc', 'hacker-news', 'job-board', 'email', 'user-supplied', 'web-search', 'other']);
@@ -208,6 +209,11 @@ function string(value, label, max = 5000) {
   return value.trim();
 }
 
+function optionalPresentString(candidate, key, max) {
+  if (!Object.hasOwn(candidate, key)) return '';
+  return string(candidate[key], `candidate.${key}`, max);
+}
+
 function stringArray(value, label, required = false) {
   if (!Array.isArray(value) || (required && value.length === 0)) throw new Error(`${label} must be ${required ? 'a non-empty' : 'an'} array of strings.`);
   return value.map((item, index) => string(item, `${label}[${index}]`, 300));
@@ -481,6 +487,26 @@ export function validateLedgerEntry(input) {
   return normalized;
 }
 
+export function validateLedgerCheckCandidate(input) {
+  const candidate = object(input, 'candidate');
+  const id = optionalPresentString(candidate, 'id', 180);
+  const url = optionalPresentString(candidate, 'url', 2048);
+  const employerJobId = optionalPresentString(candidate, 'employerJobId', 300);
+  const company = optionalPresentString(candidate, 'company', 300);
+  const role = optionalPresentString(candidate, 'role', 300);
+  if (!url && !id && !(employerJobId && company) && !(company && role)) {
+    throw new Error('ledger check requires at least one identifier set: url, id, employerJobId+company, or company+role.');
+  }
+  if (url) normalizeUrl(url);
+  return {
+    ...(id ? { id } : {}),
+    ...(url ? { url } : {}),
+    ...(employerJobId ? { employerJobId } : {}),
+    ...(company ? { company } : {}),
+    ...(role ? { role } : {}),
+  };
+}
+
 export function validateSubmissionTelemetry(input) {
   if (input == null) return {};
   const value = object(input, 'entry.telemetry');
@@ -519,6 +545,15 @@ function rolesLikelySame(left, right) {
   return shared / Math.min(leftTokens.size, rightTokens.size) >= 0.75;
 }
 
+function latestMatchingEntry(entries, predicate) {
+  const matches = entries.filter(predicate);
+  if (matches.length === 0) return null;
+  return [...matches]
+    .filter((entry) => !Number.isNaN(Date.parse(entry.submittedAt)))
+    .sort((left, right) => Date.parse(right.submittedAt) - Date.parse(left.submittedAt))[0]
+    ?? matches.at(-1);
+}
+
 const canonicalApplicationKey = accountingApplicationKey;
 
 function businessDaysBetween(startValue, endValue) {
@@ -995,21 +1030,28 @@ async function canonicalResumePath() {
 function duplicateResult(entries, candidate, outcomes = [], now = new Date()) {
   const candidateCompany = normalizedText(candidate.company);
   const candidateRole = normalizedText(candidate.role);
-  const candidateUrl = normalizeUrl(candidate.url);
+  const candidateUrl = candidate.url ? normalizeUrl(candidate.url) : null;
   const sameCompanyRole = (entry) => candidateCompany && candidateRole
     && normalizedText(entry.company) === candidateCompany
     && rolesLikelySame(entry.role, candidate.role);
-  const hardId = entries.find((entry) => entry.id === candidate.id);
+  const exactCompanyRole = (entry) => candidateCompany && candidateRole
+    && normalizedText(entry.company) === candidateCompany
+    && normalizedText(entry.role) === candidateRole;
+  const hardId = candidate.id ? entries.find((entry) => entry.id === candidate.id) : null;
   const hardEmployerJobId = entries.find((entry) => candidate.employerJobId && entry.employerJobId
     && normalizedText(entry.company) === candidateCompany
     && entry.employerJobId.toLowerCase() === String(candidate.employerJobId).toLowerCase());
-  const hardUrl = entries.find((entry) => normalizeUrl(entry.url) === candidateUrl);
+  const hardUrl = candidateUrl ? entries.find((entry) => normalizeUrl(entry.url) === candidateUrl) : null;
   const hard = hardId ?? hardEmployerJobId ?? hardUrl;
   const hardReason = hardId ? 'id' : hardEmployerJobId ? 'employer-job-id' : hardUrl ? 'url' : null;
-  const possible = hard ? null : entries.find(sameCompanyRole);
+  const exact = hard ? null : latestMatchingEntry(entries, exactCompanyRole);
+  const fuzzyMatches = hard || exact ? [] : entries.filter(sameCompanyRole);
+  const ambiguousCompanyRole = fuzzyMatches.length > 1;
+  const possible = hard ? null : exact ?? (fuzzyMatches.length === 1 ? fuzzyMatches[0] : null);
   const match = hard ?? possible;
-  const sameCompanyEntries = candidateCompany
-    ? entries.filter((entry) => normalizedText(entry.company) === candidateCompany)
+  const historyCompany = candidateCompany || normalizedText(hard?.company);
+  const sameCompanyEntries = historyCompany
+    ? entries.filter((entry) => normalizedText(entry.company) === historyCompany)
     : [];
   const companyApplications = sameCompanyEntries.slice(-20).map((entry) => ({
       id: entry.id,
@@ -1029,15 +1071,15 @@ function duplicateResult(entries, candidate, outcomes = [], now = new Date()) {
     : false;
   let companyReapplyDecision = 'fresh-company';
   if (hard) companyReapplyDecision = 'hard-duplicate';
-  else if (possible) companyReapplyDecision = 'same-role-review';
+  else if (possible || ambiguousCompanyRole) companyReapplyDecision = 'same-role-review';
   else if (latestCompanyApplication && hasFollowUp) companyReapplyDecision = 'follow-up-present';
   else if (latestCompanyApplication && daysSinceLatest < COMPANY_REAPPLY_COOLDOWN_DAYS) companyReapplyDecision = 'cooldown-active';
   else if (latestCompanyApplication) companyReapplyDecision = 'eligible-after-cooldown';
   return {
     duplicate: Boolean(hard),
-    possibleDuplicate: Boolean(possible),
-    reason: hardReason ?? (possible ? 'company-role' : null),
-    match: match ? { id: match.id, company: match.company, role: match.role, submittedAt: match.submittedAt } : null,
+    possibleDuplicate: Boolean(possible) || ambiguousCompanyRole,
+    reason: hardReason ?? (possible || ambiguousCompanyRole ? 'company-role' : null),
+    match: match ? { id: match.id, company: match.company, role: match.role, submittedAt: match.submittedAt, url: match.url } : null,
     sameCompany: companyApplications.length > 0,
     companyApplications,
     companyReapply: {
@@ -1052,12 +1094,11 @@ function duplicateResult(entries, candidate, outcomes = [], now = new Date()) {
 }
 
 async function ledgerCheck(candidate) {
-  object(candidate, 'candidate');
+  const normalized = validateLedgerCheckCandidate(candidate);
   const dir = await ensureStateDir();
   const entries = await jsonLines(join(dir, 'applications.ndjson'));
   const outcomes = await jsonLines(join(dir, 'outcomes.ndjson'));
-  string(candidate.url, 'candidate.url', 2048);
-  return duplicateResult(entries, candidate, outcomes);
+  return duplicateResult(entries, normalized, outcomes);
 }
 
 async function ledgerAdd(entryInput, duplicateOverride, companyReapplyOverride, cloudIntent = null) {
@@ -1461,7 +1502,16 @@ async function roundStatus(roundId = null) {
   const delivery = deliveryProjection(matching, await jsonLines(join(dir, 'delivery.ndjson')));
   const effectiveKeys = new Set(matching.filter((entry,i) => delivery.applications[i].counted).map(canonicalApplicationKey));
   const effectiveApplications = applications.filter(entry => effectiveKeys.has(canonicalApplicationKey(entry)));
-  const confirmedCount = delivery.effectiveSubmissionCount;
+  const outreachView = await loadOutreachRoundView(dir, cloudState);
+  const outreachCounts = projectOutreachRoundCounts({
+    applications: matching,
+    delivery,
+    roundEvents: events,
+    sentVerified: outreachView.sentVerified,
+    invalidatedIds: outreachView.invalidatedIds,
+    roundId: id,
+  });
+  const confirmedCount = outreachCounts.confirmedCount;
   const audit = started.discoveryPolicyVersion === 2 ? discoveryProjection(await jsonLines(join(dir, 'discovery.ndjson')), { roundId: id }) : null;
   const attention = await attentionList(id);
   const completion = events.find((event) => event.type === 'completed' && event.roundId === id);
@@ -1474,7 +1524,9 @@ async function roundStatus(roundId = null) {
     completed: Boolean(completion),
     discoveryPolicyVersion: started.discoveryPolicyVersion ?? 1,
     recordedSubmissionCount: delivery.recordedSubmissionCount,
-    effectiveSubmissionCount: confirmedCount,
+    applyConfirmationCount: outreachCounts.applyConfirmationCount,
+    outreachConfirmationCount: outreachCounts.outreachConfirmationCount,
+    effectiveSubmissionCount: delivery.effectiveSubmissionCount,
     failedDeliveryCount: delivery.failedDeliveryCount,
     receiptUnknownEmailCount: delivery.receiptUnknownEmailCount,
     shortfallCount: Math.max(0, started.requestedCount - confirmedCount),
@@ -1514,6 +1566,44 @@ async function roundComplete(input) {
   return result;
 }
 
+async function roundConfirm(input) {
+  const value = object(input, 'round confirmation');
+  const allowed = new Set(['roundId', 'outreachId', 'outreachIds']);
+  for (const key of Object.keys(value)) if (!allowed.has(key)) throw new Error(`Unknown round confirmation property: ${key}.`);
+  const roundId = string(value.roundId, 'round.roundId', 180);
+  const ids = [];
+  if (value.outreachId != null) ids.push(string(value.outreachId, 'round.outreachId', 100));
+  if (value.outreachIds != null) {
+    if (!Array.isArray(value.outreachIds) || value.outreachIds.length === 0 || value.outreachIds.length > 50) throw new Error('round.outreachIds must be a bounded array of outreach IDs.');
+    for (const [index, id] of value.outreachIds.entries()) ids.push(string(id, `round.outreachIds[${index}]`, 100));
+  }
+  if (!ids.length) throw new Error('round confirm requires outreachId or outreachIds.');
+  const dir = await ensureStateDir();
+  const uniqueIds = [...new Set(ids)];
+  const roundEvents = await jsonLines(join(dir, 'rounds.ndjson'));
+  const sentVerified = new Set((await loadSentVerifiedOutreach(dir, cloudState, { allowCache: false })).map((item) => item.id));
+  for (const outreachId of uniqueIds) {
+    const targetRoundId = resolveOutreachRoundId(roundEvents, { outreachId, roundId });
+    if (!targetRoundId) throw new Error('Application round was not found.');
+    const alreadyAttached = outreachConfirmationEvents(roundEvents, targetRoundId).some((event) => event.outreachId === outreachId);
+    if (!alreadyAttached && !sentVerified.has(outreachId)) throw new Error(`Outreach ${outreachId} is not sent-verified.`);
+  }
+  const confirmations = [];
+  for (const outreachId of uniqueIds) {
+    const result = await confirmOutreachTowardRound({
+      directory: dir,
+      outreachId,
+      roundId,
+      cloudClient: cloudState,
+      requireRoundId: true,
+    });
+    if (result.reason === 'round-not-found') throw new Error('Application round was not found.');
+    if (result.reason === 'not-sent-verified') throw new Error(`Outreach ${outreachId} is not sent-verified.`);
+    confirmations.push(result);
+  }
+  return { ...await roundStatus(roundId), confirmations };
+}
+
 async function frictionRecord(input) {
   const value = object(input, 'friction event');
   const allowed = new Set(['stage', 'ats', 'errorCode', 'reproducible', 'general', 'observedAt']);
@@ -1711,12 +1801,13 @@ function reviewTelemetry(review) {
 }
 
 function roundCompletedTelemetry(round) {
+  const applicationCount = round.applyConfirmationCount ?? round.effectiveSubmissionCount ?? 0;
   return {
     event: 'round_completed',
     properties: {
       requestedCount: round.requestedCount,
-      submittedCount: round.confirmedCount,
-      assessedCount: round.confirmedCount,
+      submittedCount: applicationCount,
+      assessedCount: applicationCount,
       skippedCount: 0,
       pausedCount: round.blockedCount,
       errorCount: 0,
@@ -1788,6 +1879,7 @@ async function executeCommand([area, action, value], telemetry, session, communi
     } });
   }
   else if (area === 'round' && action === 'status') result = await roundStatus(value ?? null);
+  else if (area === 'round' && action === 'confirm' && value === '--stdin') result = await roundConfirm(await jsonStdin());
   else if (area === 'round' && action === 'complete' && value === '--stdin') {
     result = await roundComplete(await jsonStdin());
     if (result.completionRecorded) domainEvents.push(roundCompletedTelemetry(result));
@@ -1813,7 +1905,7 @@ async function executeCommand([area, action, value], telemetry, session, communi
   else if (area === 'attention' && action === 'resolve' && value === '--stdin') result = await attentionResolve(await jsonStdin());
   else if (area === 'friction' && action === 'record' && value === '--stdin') result = await frictionRecord(await jsonStdin());
   else if (area === 'friction' && action === 'list' && value == null) result = await frictionList();
-  else throw new Error('Usage: cloud status|configure --stdin|reconcile [--dry-run]|export [path]|lease-acquire|lease-renew|lease-release|intent-prepare --stdin|intent-sent --stdin|intent-confirm --stdin; profile set|migrate --stdin; profile check|field <name>; resume import <url-or-pdf>|path; score --stdin; ledger check|add|outcome|review-ack --stdin; ledger review; autonomy grant --stdin|status|preview|revoke; round start|source|complete --stdin|status [round-id]; sources list [--stdin]|jobs [--stdin]|suggest --stdin|pending|sync|sharing status|enable|disable|reset; attention add|resolve --stdin|list; friction record --stdin|list; telemetry status|enable|disable|reset|preview --stdin|record --stdin');
+  else throw new Error('Usage: cloud status|configure --stdin|reconcile [--dry-run]|export [path]|lease-acquire|lease-renew|lease-release|intent-prepare --stdin|intent-sent --stdin|intent-confirm --stdin; profile set|migrate --stdin; profile check|field <name>; resume import <url-or-pdf>|path; score --stdin; ledger check|add|outcome|review-ack --stdin; ledger review; autonomy grant --stdin|status|preview|revoke; round start|source|confirm|complete --stdin|status [round-id]; sources list [--stdin]|jobs [--stdin]|suggest --stdin|pending|sync|sharing status|enable|disable|reset; attention add|resolve --stdin|list; friction record --stdin|list; telemetry status|enable|disable|reset|preview --stdin|record --stdin');
   for (const event of domainEvents) await telemetry.record(event, session);
   return result;
 }

package/job-application-agent/scripts/outreach-cli.mjs

@@ -5,6 +5,7 @@ import { join } from 'node:path';
 import { CloudStateClient, defaultCloudConfigPath } from './cloud-state-client.mjs';
 import { migrateLegacyStateDir, resolveStateDir } from './secret-store.mjs';
 import { mutateOutreach, readOutreach, OUTREACH_CAPABILITY } from './outreach-domain.mjs';
+import { confirmOutreachTowardRound } from './outreach-round.mjs';
 import { privateOutreachWrite, withLocalOutreach, withOutreachLock } from './outreach-store.mjs';
 
 const READS = new Set(['policy-status', 'list', 'show', 'review']);
@@ -91,5 +92,21 @@ export async function runOutreach(args, { input: suppliedInput, stateDirectory =
     result = await withLocalOutreach(stateDirectory, state => read ? { result: readOutreach(state, action, input) } : mutateOutreach(state, action, input, { ...context, actor: 'local' }));
   }
   if (action === 'policy-enable') await guard();
+  if (action === 'record' && input?.type === 'sent-verified' && result?.delivery === 'sent-verified' && input?.id) {
+    try {
+      result = {
+        ...result,
+        roundConfirmation: await confirmOutreachTowardRound({
+          directory: stateDirectory,
+          outreachId: input.id,
+          cloudClient: cloud,
+          occurredAt: input.occurredAt,
+        }),
+      };
+    } catch {
+      // The outreach observation already committed; attachment is a secondary result.
+      result = { ...result, roundConfirmation: { counted: false, reason: 'attachment-failed', outreachId: input.id } };
+    }
+  }
   return result;
 }