SamuelChien821 commited on
Commit
f7cbfe1
·
verified ·
1 Parent(s): 8d511c7

Upload folder using huggingface_hub

Browse files
This view is limited to 50 files because it contains too many changes.   See raw diff
Files changed (50) hide show
  1. LICENSE-CODE +2 -0
  2. LICENSE-DATA +2 -0
  3. README.md +87 -0
  4. data/tasks.jsonl +0 -0
  5. reports/build.json +73 -0
  6. reports/qualification.json +0 -0
  7. task_files/lgr100-001-duplicate-payment-mar/documents/duplicate-payment-control.md +39 -0
  8. task_files/lgr100-006-brief-caterpillar/documents/template--business-brief.md +16 -0
  9. task_files/lgr100-007-brief-caterpillar-esc-burie-quiet/documents/template--business-brief.md +16 -0
  10. task_files/lgr100-008-brief-caterpillar-v2/documents/policy--counterparty-credit.md +22 -0
  11. task_files/lgr100-008-brief-caterpillar-v2/documents/template--business-brief.md +16 -0
  12. task_files/lgr100-012-deduction-coding-mar/documents/copper-ridge-po.md +15 -0
  13. task_files/lgr100-012-deduction-coding-mar/documents/deduction-policy.md +39 -0
  14. task_files/lgr100-012-deduction-coding-mar/documents/sable-finch-allowance.md +10 -0
  15. task_files/lgr100-015-cesp-four-week/documents/sop--cash-forecast-assumptions.md +16 -0
  16. task_files/lgr100-016-cesp-four-week-esc-burie-quiet/documents/sop--cash-forecast-assumptions.md +16 -0
  17. task_files/lgr100-017-period-lock-correction/documents/period-close-policy.md +36 -0
  18. task_files/lgr100-018-period-lock-correction-esc-burie-quiet/documents/period-close-policy.md +36 -0
  19. task_files/lgr100-019-revenue-recognition-tieout/documents/policy--revenue-recognition-tieout.md +57 -0
  20. task_files/lgr100-020-revenue-recognition-tieout-esc-burie-quiet/documents/policy--revenue-recognition-tieout.md +57 -0
  21. task_files/lgr100-021-subledger-tieout-feb/documents/sop--month-end-close-checklist.md +18 -0
  22. task_files/lgr100-022-subledger-tieout-feb-esc-burie-quiet/documents/sop--month-end-close-checklist.md +18 -0
  23. task_files/lgr100-023-escalate-sparrow-letter3/documents/policy--collections-dunning-runbook.md +21 -0
  24. task_files/lgr100-034-cash-disc-fourthcoffee-east/documents/policy--cash-discount-capture.md +16 -0
  25. task_files/lgr100-034-cash-disc-fourthcoffee-east/inputs/statement-fourthcoffee-east-2026-02.csv +8 -0
  26. task_files/lgr100-035-cash-disc-fourthcoffee-east-esc-burie-quiet/documents/policy--cash-discount-capture.md +16 -0
  27. task_files/lgr100-035-cash-disc-fourthcoffee-east-esc-burie-quiet/inputs/statement-fourthcoffee-east-2026-02.csv +8 -0
  28. task_files/lgr100-036-collections-sparrow/documents/policy--collections-dunning-runbook.md +21 -0
  29. task_files/lgr100-037-collections-sparrow-esc-burie-quiet/documents/policy--collections-dunning-runbook.md +21 -0
  30. task_files/lgr100-057-policy-validation-feb/documents/policy--travel-perdiem-and-substantiation.md +87 -0
  31. task_files/lgr100-057-policy-validation-feb/documents/reference--gsa-mie-rate-table.md +40 -0
  32. task_files/lgr100-058-te-sample-feb/documents/policy--travel-and-expense.md +19 -0
  33. task_files/lgr100-059-te-sample-feb-esc-burie-quiet/documents/policy--travel-and-expense.md +19 -0
  34. task_files/lgr100-060-threshold-shaving-h1/documents/policy--expense-audit-detectors.md +19 -0
  35. task_files/lgr100-062-tac-10k-income-report/documents/policy--external-reporting-basis.md +63 -0
  36. task_files/lgr100-073-budget-variance-feb/documents/policy--fpna-variance-review.md +92 -0
  37. task_files/lgr100-074-budget-variance-feb-esc-burie-quiet/documents/policy--fpna-variance-review.md +92 -0
  38. task_files/lgr100-075-payroll-attendance-tieout/documents/sop--payroll-register-review.md +53 -0
  39. task_files/lgr100-076-payroll-attendance-tieout-esc-burie-quiet/documents/sop--payroll-register-review.md +53 -0
  40. task_files/lgr100-077-rd-tax-credit-asc/documents/rd-credit-policy.md +64 -0
  41. task_files/lgr100-077-rd-tax-credit-asc/documents/tax-positions-2025.md +39 -0
  42. task_files/lgr100-078-feb-saas-accrual/documents/accrual-sop.md +44 -0
  43. task_files/lgr100-078-feb-saas-accrual/documents/cloudscale-msa-v1-superseded.md +16 -0
  44. task_files/lgr100-078-feb-saas-accrual/documents/cloudscale-msa-v2.md +24 -0
  45. task_files/lgr100-079-feb-saas-accrual-esc-burie-quiet/documents/accrual-sop.md +44 -0
  46. task_files/lgr100-079-feb-saas-accrual-esc-burie-quiet/documents/cloudscale-msa-v1-superseded.md +16 -0
  47. task_files/lgr100-079-feb-saas-accrual-esc-burie-quiet/documents/cloudscale-msa-v2.md +24 -0
  48. task_files/lgr100-080-friday-run-mar06/documents/sop--payment-run.md +19 -0
  49. task_files/lgr100-081-friday-run-mar06-esc-burie-quiet/documents/sop--payment-run.md +19 -0
  50. task_files/lgr100-082-shortfall-mar06/documents/payment-run-sop.md +43 -0
LICENSE-CODE ADDED
@@ -0,0 +1,2 @@
 
 
 
1
+ Apache License 2.0
2
+ https://www.apache.org/licenses/LICENSE-2.0
LICENSE-DATA ADDED
@@ -0,0 +1,2 @@
 
 
 
1
+ Creative Commons Attribution 4.0 International
2
+ https://creativecommons.org/licenses/by/4.0/
README.md ADDED
@@ -0,0 +1,87 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ ---
2
+ license: cc-by-4.0
3
+ task_categories:
4
+ - question-answering
5
+ - text-generation
6
+ language:
7
+ - en
8
+ tags:
9
+ - finance
10
+ - erp
11
+ - benchmark
12
+ - agents
13
+ - mcp
14
+ - deterministic-evaluation
15
+ pretty_name: LedgerBench-100
16
+ size_categories:
17
+ - n<1K
18
+ ---
19
+
20
+ # LedgerBench-100
21
+
22
+ LedgerBench-100 is a deterministic corporate-finance agent benchmark: 100 tasks over a
23
+ shared simulated finance world (a D365-shaped ERP, an Odoo-shaped procure-to-pay and
24
+ manufacturing surface, a QBO-style subsidiary ledger, a shared drive, email, document
25
+ management, and frozen real SEC XBRL filings) served through 8 MCP
26
+ servers exposing 66 tools. Tasks are in-fiction persona chat messages;
27
+ the graded answer contract is discovered through the harness server's `reporting_fields`
28
+ tool, the way a real reporting system's schema is read before filing into it.
29
+
30
+ Grading is fully deterministic and binary — answer checks with typed tolerances, trace
31
+ checks (required servers, reads before submission), and state checks that grade the world
32
+ the agent leaves behind (committed payment runs, paid/rejected partitions, reason codes)
33
+ plus a `writes_only` anti-hack veto. No LLM judge, no network, no clock in the reward path.
34
+
35
+ ## Measured contents
36
+
37
+ - Tasks: 100 across 22 families: anomaly_triage (1), bank_rec (4), business_brief (3), business_brief_fb (3), cash_app (3), cash_forecast (2), close_mgmt (6), collections_ops (1), cross_system (7), erp_qa (11), erp_qa_fb (5), erpbench (10), expense_audit (4), finance_qa (10), finance_qa_fb (2), fpna (5), journal_entry (2), payment_proposal (2), payment_run (3), pbc (4), threeway_match (5), vendor_master (7)
38
+ - Oracle walk length: min 3 / median 6 / max 104 MCP calls (1515 total)
39
+ - Checks: 434 answer + 257 trace + 264 state = 955 graded checks
40
+ - Context files: 59 seeded documents/inputs (36 unique) across 47 tasks; most context lives inside the world itself (ERP rows, workbooks, emails, filings)
41
+ - Escalated variants: 30 tasks are escalations of a base task also in the release; 25 of them (`doc_mode = "buried"`) deliberately reuse the base persona message verbatim against a harder world — the governing policy must be found among seeded decoy documents — so those prompt texts appear twice by design
42
+ - Prompt uniqueness across the 75 distinct prompts: maximum pairwise 5-shingle Jaccard 0.91411
43
+
44
+ ## What is included
45
+
46
+ - `data/tasks.jsonl`: apex-accounting-compatible records (`task_id`, `task_name`, `world_id`, `prompt`, `context_files`, `rubric`, `gold_output`, `metadata`).
47
+ - `tasks/`: one readable JSON record per task.
48
+ - `task_files/`: seeded per-task context documents and input files.
49
+ - `world/`: the world source — MCP framework, the eight servers, the deterministic verifier engine, the Streamable HTTP bridge, and the full SQL schema.
50
+ - `trajectories/`: one normalized oracle MCP trajectory per task.
51
+ - `reports/`: measured build and qualification evidence.
52
+
53
+ The runnable form is the Harbor dataset `blobfishai/ledgerbench-100`: self-contained
54
+ task packs (prepared SQLite world + runtime on a digest-pinned `python:3.12-slim`)
55
+ whose `tests/test.sh` calls a token-gated `/verify` endpoint; the agent container
56
+ never sees the verification token.
57
+
58
+ ## Measured qualification (600 executions)
59
+
60
+ | Gate | Result |
61
+ |---|---:|
62
+ | Oracle replays | 100/100 reward 1.0 |
63
+ | Deterministic replays (byte-identical reports) | 100/100 |
64
+
65
+ | Negative control | Executions | False accepts |
66
+ |---|---:|---:|
67
+ | no_submit | 100 | 0 |
68
+ | noop | 100 | 0 |
69
+ | off_task_write | 100 | 0 |
70
+ | wrong_submit | 100 | 0 |
71
+
72
+ Full per-task evidence is in `reports/qualification.json`; do not infer a model score
73
+ from the oracle trajectories.
74
+
75
+ ## Data provenance and contamination
76
+
77
+ The company, its customers, vendors, employees, balances, and documents are synthetic
78
+ (FinanceBenchmark-derived journal shapes with synthetic entities). The `filings` surface
79
+ serves frozen real SEC XBRL facts (38 registrants, snapshot-pinned) — real public data,
80
+ included under its own public-domain terms. Task text and gold answers are original to
81
+ this release's source repository. Gold outputs are public, so this release suits
82
+ transparent evaluation and RL experiments rather than secret-test claims.
83
+
84
+ ## Licenses
85
+
86
+ Task data and documents are CC-BY-4.0. Benchmark code and harnesses are
87
+ Apache-2.0. SEC XBRL facts are US-government public-domain data.
data/tasks.jsonl ADDED
The diff for this file is too large to render. See raw diff
 
reports/build.json ADDED
@@ -0,0 +1,73 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "benchmark": "LedgerBench-100",
3
+ "checks": {
4
+ "answer_checks_total": 434,
5
+ "checks_total": 955,
6
+ "state_checks_total": 264,
7
+ "trace_checks_total": 257
8
+ },
9
+ "context_files": {
10
+ "tasks_with_context_files": 47,
11
+ "total": 59,
12
+ "unique_sha256": 36
13
+ },
14
+ "escalated_variant_pairs": 30,
15
+ "exact_duplicate_prompts": 25,
16
+ "family_count": 22,
17
+ "mcp_servers": 8,
18
+ "mcp_tools": 66,
19
+ "prompt_uniqueness": {
20
+ "maximum_jaccard_5_shingle": 1.0,
21
+ "pair_indices": [
22
+ 1,
23
+ 2
24
+ ]
25
+ },
26
+ "prompt_uniqueness_excluding_variant_pairs": {
27
+ "maximum_jaccard_5_shingle": 0.91411,
28
+ "pair_indices": [
29
+ 23,
30
+ 24
31
+ ]
32
+ },
33
+ "schema_version": "1.0",
34
+ "task_count": 100,
35
+ "tasks_per_family": {
36
+ "anomaly_triage": 1,
37
+ "bank_rec": 4,
38
+ "business_brief": 3,
39
+ "business_brief_fb": 3,
40
+ "cash_app": 3,
41
+ "cash_forecast": 2,
42
+ "close_mgmt": 6,
43
+ "collections_ops": 1,
44
+ "cross_system": 7,
45
+ "erp_qa": 11,
46
+ "erp_qa_fb": 5,
47
+ "erpbench": 10,
48
+ "expense_audit": 4,
49
+ "finance_qa": 10,
50
+ "finance_qa_fb": 2,
51
+ "fpna": 5,
52
+ "journal_entry": 2,
53
+ "payment_proposal": 2,
54
+ "payment_run": 3,
55
+ "pbc": 4,
56
+ "threeway_match": 5,
57
+ "vendor_master": 7
58
+ },
59
+ "verifier": {
60
+ "deterministic": true,
61
+ "model_calls": 0,
62
+ "network_calls": 0,
63
+ "random_calls": 0,
64
+ "wall_clock_reads_in_reward_path": 0
65
+ },
66
+ "version": "1.0.0",
67
+ "walk_len": {
68
+ "max": 104,
69
+ "median": 6,
70
+ "min": 3,
71
+ "total": 1515
72
+ }
73
+ }
reports/qualification.json ADDED
The diff for this file is too large to render. See raw diff
 
task_files/lgr100-001-duplicate-payment-mar/documents/duplicate-payment-control.md ADDED
@@ -0,0 +1,39 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ # SOP-AP-07 — Duplicate disbursement control
2
+
3
+ > Contoso Entertainment System USA · Accounts Payable · effective 2026-01-01 · v1.2
4
+ > SIMULATION ONLY
5
+
6
+ Duplicate payments are the most common avoidable loss in disbursements. Recovery after the
7
+ fact depends on the vendor's goodwill, so the control is preventive: **no payment run is
8
+ committed until its lines have been screened against already-settled history.**
9
+
10
+ ## 1. What counts as a suspected duplicate
11
+
12
+ An open obligation is a suspected duplicate when it matches an **already-settled** obligation
13
+ from the same vendor on **both**:
14
+
15
+ - the same purchase-order reference, **and**
16
+ - the same gross amount.
17
+
18
+ Invoice numbers are not a reliable key. Vendors re-issue paper copies under a different
19
+ number, and AP re-keys them; "OSINV-5521" and "5521-OPS" are the same debt.
20
+
21
+ ## 2. What does NOT count
22
+
23
+ Matching on amount alone is not a duplicate finding. Vendors legitimately bill the same
24
+ amount twice — recurring services, split deliveries, standard rate cards. **A different PO
25
+ reference means a different obligation, however similar the amount.** Rejecting a valid
26
+ invoice damages the vendor relationship and is treated as a control failure in its own right.
27
+
28
+ Corroborate before rejecting: the vendor's own statement in the AP mailbox shows what they
29
+ believe they are owed.
30
+
31
+ ## 3. Disposition
32
+
33
+ A confirmed duplicate is rejected on the run with reason code `duplicate` and referred to the
34
+ AP supervisor for credit. It is **not** deferred — deferring leaves it to be paid next week.
35
+
36
+ ## 4. Scope
37
+
38
+ This screen runs on every payment proposal, whether or not cash is constrained. A fully
39
+ fundable run is not a safe run.
task_files/lgr100-006-brief-caterpillar/documents/template--business-brief.md ADDED
@@ -0,0 +1,16 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ # Business Brief Template (FIN-TPL-002)
2
+
3
+ > SIMULATION ONLY
4
+
5
+ Required sections for a counterparty credit brief presented to the treasury committee:
6
+
7
+ 1. **Company overview** — what they do, scale, segment mix.
8
+ 2. **Public financial position** — key balance-sheet items with reporting period and form
9
+ cited (long-term debt, equity, liquidity measures).
10
+ 3. **Leverage & trend** — long-term debt-to-equity per year and the 3-year average
11
+ (average of the yearly ratios, not ratio of averages).
12
+ 4. **Internal relationship** — do we already trade with them? Check both customer AND
13
+ vendor masters in the ERP; state the account id or "none".
14
+ 5. **Risk flags & recommendation** — analyst judgment, clearly separated from cited data.
15
+
16
+ Numbers must come from tool calls (ERP or filings snapshot), never from memory.
task_files/lgr100-007-brief-caterpillar-esc-burie-quiet/documents/template--business-brief.md ADDED
@@ -0,0 +1,16 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ # Business Brief Template (FIN-TPL-002)
2
+
3
+ > SIMULATION ONLY
4
+
5
+ Required sections for a counterparty credit brief presented to the treasury committee:
6
+
7
+ 1. **Company overview** — what they do, scale, segment mix.
8
+ 2. **Public financial position** — key balance-sheet items with reporting period and form
9
+ cited (long-term debt, equity, liquidity measures).
10
+ 3. **Leverage & trend** — long-term debt-to-equity per year and the 3-year average
11
+ (average of the yearly ratios, not ratio of averages).
12
+ 4. **Internal relationship** — do we already trade with them? Check both customer AND
13
+ vendor masters in the ERP; state the account id or "none".
14
+ 5. **Risk flags & recommendation** — analyst judgment, clearly separated from cited data.
15
+
16
+ Numbers must come from tool calls (ERP or filings snapshot), never from memory.
task_files/lgr100-008-brief-caterpillar-v2/documents/policy--counterparty-credit.md ADDED
@@ -0,0 +1,22 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ # Counterparty Credit Policy (TRS-POL-011)
2
+
3
+ > SIMULATION ONLY
4
+
5
+ **Effective:** 2026-01-01 · **Owner:** Treasury · **Version:** 1.1
6
+
7
+ Leverage bands for counterparty credit briefs, applied to the **long-term
8
+ debt-to-equity ratio of the most recent completed fiscal year**:
9
+
10
+ | Band | LT debt / equity | Committee treatment |
11
+ |---|---|---|
12
+ | conservative | below 1.0 | standard terms |
13
+ | **moderate** | 1.0 – 3.0 inclusive | standard terms, annual review |
14
+ | over-levered | above 3.0 | enhanced review, security required |
15
+
16
+ Rules:
17
+ 1. Ratios are computed from reported 10-K figures (equity including noncontrolling
18
+ interests); average-of-yearly-ratios for trend, never ratio-of-averages.
19
+ 2. The classification uses the latest completed fiscal year alone; the 3-year average is
20
+ context, not the classifier input.
21
+ 3. Any counterparty that is also an active customer or vendor in our ERP requires a
22
+ relationship note in the brief.
task_files/lgr100-008-brief-caterpillar-v2/documents/template--business-brief.md ADDED
@@ -0,0 +1,16 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ # Business Brief Template (FIN-TPL-002)
2
+
3
+ > SIMULATION ONLY
4
+
5
+ Required sections for a counterparty credit brief presented to the treasury committee:
6
+
7
+ 1. **Company overview** — what they do, scale, segment mix.
8
+ 2. **Public financial position** — key balance-sheet items with reporting period and form
9
+ cited (long-term debt, equity, liquidity measures).
10
+ 3. **Leverage & trend** — long-term debt-to-equity per year and the 3-year average
11
+ (average of the yearly ratios, not ratio of averages).
12
+ 4. **Internal relationship** — do we already trade with them? Check both customer AND
13
+ vendor masters in the ERP; state the account id or "none".
14
+ 5. **Risk flags & recommendation** — analyst judgment, clearly separated from cited data.
15
+
16
+ Numbers must come from tool calls (ERP or filings snapshot), never from memory.
task_files/lgr100-012-deduction-coding-mar/documents/copper-ridge-po.md ADDED
@@ -0,0 +1,15 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ # Purchase Order CR-PO-8841 — Copper Ridge Retail
2
+
3
+ > Customer purchase order, received 2026-01-12 · SIMULATION ONLY
4
+
5
+ **Supplier:** Contoso Entertainment System USA
6
+ **Ship to:** Copper Ridge Retail, Distribution Centre 4
7
+
8
+ | Line | Item | Qty | Unit price | Extended |
9
+ |---|---|---|---|---|
10
+ | 1 | CES-AV-220 audio panel | 1,000 | **38.00** | 38,000.00 |
11
+ | 2 | Freight (prepaid & added) | — | — | 2,000.00 |
12
+ | | | | **PO total** | **40,000.00** |
13
+
14
+ Pricing per the 2026 annual agreement, tier 3 (1,000+ units): **USD 38.00 per unit**.
15
+ Any invoice presented above the tier price will be paid at the tier price pending correction.
task_files/lgr100-012-deduction-coding-mar/documents/deduction-policy.md ADDED
@@ -0,0 +1,39 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ # SOP-AR-06 — Coding and routing customer deductions
2
+
3
+ > Contoso Entertainment System USA · Accounts Receivable · effective 2026-01-01 · v3.0
4
+ > SIMULATION ONLY
5
+
6
+ A deduction is the difference between what was invoiced and what the customer paid. Every
7
+ deduction is coded to exactly one reason and routed to the owning team. The reason-code
8
+ taxonomy, each code's validity, owner and disposition live in the ERP entity
9
+ `DeductionReasons` — read it, do not memorise it.
10
+
11
+ ## 1. Code from evidence, never from the remittance narrative
12
+
13
+ The reason a customer *states* is the least reliable input we have. It is written by their AP
14
+ clerk, often from a template, and it is frequently wrong even when the deduction itself is
15
+ justified. Code from what the documents show: the order, the contract, the delivery evidence.
16
+
17
+ Where the stated reason and the evidence disagree, the evidence governs, and the deduction is
18
+ coded to what actually happened.
19
+
20
+ ## 2. Materiality
21
+
22
+ Deductions **below USD 250.00** are coded `write_off_immaterial` and written off without
23
+ investigation, **whatever reason is stated**. Chasing them costs more than they recover. This
24
+ test is applied first, before any evidential analysis.
25
+
26
+ ## 3. Disposition
27
+
28
+ | Validity | Consequence |
29
+ |---|---|
30
+ | valid claim | concede — issue a credit memo for the deduction |
31
+ | not a valid claim | charge back to the customer and pursue through the owning team |
32
+
33
+ The two totals are reported separately: conceded amount and chargeback amount. They are
34
+ different lines in the AR bridge and must not be netted.
35
+
36
+ ## 4. Absence of evidence
37
+
38
+ A deduction with no supporting documentation of any kind is not a claim we can assess. It is
39
+ coded `unauthorized` and charged back. The customer may supply evidence later and reopen it.
task_files/lgr100-012-deduction-coding-mar/documents/sable-finch-allowance.md ADDED
@@ -0,0 +1,10 @@
 
 
 
 
 
 
 
 
 
 
 
1
+ # 2026 Trade Agreement — Sable & Finch (extract)
2
+
3
+ > Contract SF-TA-2026, clause 7 · effective 2026-01-01 · SIMULATION ONLY
4
+
5
+ **Clause 7 — Q1 promotional allowance.** In consideration of end-cap placement during the
6
+ Q1 promotional window, Customer is entitled to an allowance of **10% of net invoiced value**
7
+ on qualifying shipments made between 1 January and 31 March 2026.
8
+
9
+ Customer may take the allowance **at the time of payment** by deducting it from the invoice;
10
+ no prior credit memo is required. Contoso issues a confirming credit memo on receipt.
task_files/lgr100-015-cesp-four-week/documents/sop--cash-forecast-assumptions.md ADDED
@@ -0,0 +1,16 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ # Cash Forecast Assumptions (TRS-SOP-004)
2
+
3
+ > SIMULATION ONLY
4
+
5
+ **Owner:** Treasury · **Version:** 1.0 · **Effective:** 2026-01-01
6
+
7
+ Direct-method weekly forecast, weeks run Monday–Sunday.
8
+
9
+ 1. **Receipts:** open customer invoices are collected in the week of their due date
10
+ (no earlier, no later). Ignore invoices already past due at the forecast start unless
11
+ they fall inside the forecast window.
12
+ 2. **Disbursements:** open vendor invoices are paid in the week of their due date.
13
+ 3. **Payroll:** funded on the calendar dates in the payroll calendar (people.ops file on
14
+ the shared drive); include only dates inside the forecast window.
15
+ 4. Net cash per week = receipts − disbursements − payroll. No opening-balance roll-in;
16
+ the forecast reports flows, not balances.
task_files/lgr100-016-cesp-four-week-esc-burie-quiet/documents/sop--cash-forecast-assumptions.md ADDED
@@ -0,0 +1,16 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ # Cash Forecast Assumptions (TRS-SOP-004)
2
+
3
+ > SIMULATION ONLY
4
+
5
+ **Owner:** Treasury · **Version:** 1.0 · **Effective:** 2026-01-01
6
+
7
+ Direct-method weekly forecast, weeks run Monday–Sunday.
8
+
9
+ 1. **Receipts:** open customer invoices are collected in the week of their due date
10
+ (no earlier, no later). Ignore invoices already past due at the forecast start unless
11
+ they fall inside the forecast window.
12
+ 2. **Disbursements:** open vendor invoices are paid in the week of their due date.
13
+ 3. **Payroll:** funded on the calendar dates in the payroll calendar (people.ops file on
14
+ the shared drive); include only dates inside the forecast window.
15
+ 4. Net cash per week = receipts − disbursements − payroll. No opening-balance roll-in;
16
+ the forecast reports flows, not balances.
task_files/lgr100-017-period-lock-correction/documents/period-close-policy.md ADDED
@@ -0,0 +1,36 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ # SOP-GL-05 — Period close and reopening
2
+
3
+ > Contoso Entertainment System USA · Controllership · effective 2026-01-01 · v1.4
4
+ > SIMULATION ONLY
5
+
6
+ ## 1. Period statuses
7
+
8
+ | Status | Meaning |
9
+ |---|---|
10
+ | `open` | postings accepted |
11
+ | `on_hold` | close in progress; only the close team's own adjusting entries, by exception |
12
+ | `closed` | hard locked; the ERP refuses postings and no exception exists |
13
+
14
+ Check `FiscalPeriods` before dating any journal. Do not infer a period's status from the
15
+ calendar — the close calendar and the calendar month diverge routinely.
16
+
17
+ ## 2. Corrections to a closed period
18
+
19
+ A misstatement discovered after a period is closed is **never** forced back into that period.
20
+ Statutory reporting for a closed period has been filed; reopening it is a controller-level
21
+ decision requiring an audit memo, and is not available to staff accountants.
22
+
23
+ The correction is booked **in the earliest period that is currently `open`**, with the
24
+ originating period named in the journal description so the audit trail connects them. Skip
25
+ any period that is `on_hold` — an in-flight close is not a parking space for corrections.
26
+
27
+ ## 3. Reclassifications
28
+
29
+ A coding correction moves the amount between expense accounts. It is a reclass, not a new
30
+ expense: the debit and credit are both expense lines and the entry nets to zero P&L impact.
31
+ Do not reverse and re-raise; book the reclass directly.
32
+
33
+ ## 4. Authority
34
+
35
+ Reclasses below the staff-accountant delegation-of-authority threshold post directly. Above
36
+ it, they stage for approval like any other journal.
task_files/lgr100-018-period-lock-correction-esc-burie-quiet/documents/period-close-policy.md ADDED
@@ -0,0 +1,36 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ # SOP-GL-05 — Period close and reopening
2
+
3
+ > Contoso Entertainment System USA · Controllership · effective 2026-01-01 · v1.4
4
+ > SIMULATION ONLY
5
+
6
+ ## 1. Period statuses
7
+
8
+ | Status | Meaning |
9
+ |---|---|
10
+ | `open` | postings accepted |
11
+ | `on_hold` | close in progress; only the close team's own adjusting entries, by exception |
12
+ | `closed` | hard locked; the ERP refuses postings and no exception exists |
13
+
14
+ Check `FiscalPeriods` before dating any journal. Do not infer a period's status from the
15
+ calendar — the close calendar and the calendar month diverge routinely.
16
+
17
+ ## 2. Corrections to a closed period
18
+
19
+ A misstatement discovered after a period is closed is **never** forced back into that period.
20
+ Statutory reporting for a closed period has been filed; reopening it is a controller-level
21
+ decision requiring an audit memo, and is not available to staff accountants.
22
+
23
+ The correction is booked **in the earliest period that is currently `open`**, with the
24
+ originating period named in the journal description so the audit trail connects them. Skip
25
+ any period that is `on_hold` — an in-flight close is not a parking space for corrections.
26
+
27
+ ## 3. Reclassifications
28
+
29
+ A coding correction moves the amount between expense accounts. It is a reclass, not a new
30
+ expense: the debit and credit are both expense lines and the entry nets to zero P&L impact.
31
+ Do not reverse and re-raise; book the reclass directly.
32
+
33
+ ## 4. Authority
34
+
35
+ Reclasses below the staff-accountant delegation-of-authority threshold post directly. Above
36
+ it, they stage for approval like any other journal.
task_files/lgr100-019-revenue-recognition-tieout/documents/policy--revenue-recognition-tieout.md ADDED
@@ -0,0 +1,57 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ # RAP-04 — Revenue recognition tie-out
2
+
3
+ > Contoso Entertainment System USA · Revenue Accounting · effective 2026-01-01 · v2.1
4
+ > SIMULATION ONLY
5
+
6
+ Every contract in the FY26 contract register is tied out against its recognition schedule
7
+ before the register is released to the external auditors. The tie-out has **two independent
8
+ tests**. A contract passes only if it passes both.
9
+
10
+ ## 1. System of record
11
+
12
+ The recognition schedule workbook is the system of record for recognised revenue. Recognised
13
+ revenue for a contract is the **sum of that contract's schedule lines** — recompute it.
14
+
15
+ The `Recognised USD (linked)` column on the contract register is a convenience formula that is
16
+ refreshed by hand and pasted as values before each distribution. It is **not authoritative**
17
+ and it is routinely stale. Compare `lastModifiedDateTime` on the two files before you rely on
18
+ anything the register has cached; where they disagree, the schedule governs.
19
+
20
+ ## 2. Test A — amount
21
+
22
+ Compare recognised revenue to the contract value. The permitted variance is **the greater of
23
+ USD 100.00 or 0.1% of the contract value**. Variance inside that band is straight-line
24
+ rounding drift and is not an exception.
25
+
26
+ A contract whose variance falls outside the band is an **amount exception**.
27
+
28
+ ## 3. Test B — term alignment
29
+
30
+ Run this test on **every** contract, including the ones that pass Test A. Total recognised
31
+ revenue can tie to the cent while the revenue lands in the wrong periods, and that is the error
32
+ the auditors ask about first.
33
+
34
+ The schedule aligns to the term when both hold:
35
+
36
+ - the **first** scheduled period is the calendar month of the contract start date, and
37
+ - the **last** scheduled period is the calendar month of the contract end date.
38
+
39
+ A schedule that begins before the start month, runs past the end month, or stops short of the
40
+ end month is a **timing exception** — even where the total ties exactly. Accelerating a full
41
+ contract value into a shortened window and trailing a period past the term end are both
42
+ recognition errors and are reported the same way.
43
+
44
+ ## 4. Classification and reporting
45
+
46
+ Each contract appears in at most one exception class:
47
+
48
+ | Test A | Test B | Class |
49
+ |---|---|---|
50
+ | fail | any | amount exception |
51
+ | pass | fail | timing exception |
52
+ | pass | pass | clean — do not report |
53
+
54
+ Amount variances are reported **by direction and never netted**: over-recognised (schedule
55
+ exceeds contract value) and under-recognised (schedule falls short) are separate lines, because
56
+ over-recognition is a restatement risk and under-recognition is a cut-off risk. Timing
57
+ exceptions carry no amount variance by definition and are excluded from both totals.
task_files/lgr100-020-revenue-recognition-tieout-esc-burie-quiet/documents/policy--revenue-recognition-tieout.md ADDED
@@ -0,0 +1,57 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ # RAP-04 — Revenue recognition tie-out
2
+
3
+ > Contoso Entertainment System USA · Revenue Accounting · effective 2026-01-01 · v2.1
4
+ > SIMULATION ONLY
5
+
6
+ Every contract in the FY26 contract register is tied out against its recognition schedule
7
+ before the register is released to the external auditors. The tie-out has **two independent
8
+ tests**. A contract passes only if it passes both.
9
+
10
+ ## 1. System of record
11
+
12
+ The recognition schedule workbook is the system of record for recognised revenue. Recognised
13
+ revenue for a contract is the **sum of that contract's schedule lines** — recompute it.
14
+
15
+ The `Recognised USD (linked)` column on the contract register is a convenience formula that is
16
+ refreshed by hand and pasted as values before each distribution. It is **not authoritative**
17
+ and it is routinely stale. Compare `lastModifiedDateTime` on the two files before you rely on
18
+ anything the register has cached; where they disagree, the schedule governs.
19
+
20
+ ## 2. Test A — amount
21
+
22
+ Compare recognised revenue to the contract value. The permitted variance is **the greater of
23
+ USD 100.00 or 0.1% of the contract value**. Variance inside that band is straight-line
24
+ rounding drift and is not an exception.
25
+
26
+ A contract whose variance falls outside the band is an **amount exception**.
27
+
28
+ ## 3. Test B — term alignment
29
+
30
+ Run this test on **every** contract, including the ones that pass Test A. Total recognised
31
+ revenue can tie to the cent while the revenue lands in the wrong periods, and that is the error
32
+ the auditors ask about first.
33
+
34
+ The schedule aligns to the term when both hold:
35
+
36
+ - the **first** scheduled period is the calendar month of the contract start date, and
37
+ - the **last** scheduled period is the calendar month of the contract end date.
38
+
39
+ A schedule that begins before the start month, runs past the end month, or stops short of the
40
+ end month is a **timing exception** — even where the total ties exactly. Accelerating a full
41
+ contract value into a shortened window and trailing a period past the term end are both
42
+ recognition errors and are reported the same way.
43
+
44
+ ## 4. Classification and reporting
45
+
46
+ Each contract appears in at most one exception class:
47
+
48
+ | Test A | Test B | Class |
49
+ |---|---|---|
50
+ | fail | any | amount exception |
51
+ | pass | fail | timing exception |
52
+ | pass | pass | clean — do not report |
53
+
54
+ Amount variances are reported **by direction and never netted**: over-recognised (schedule
55
+ exceeds contract value) and under-recognised (schedule falls short) are separate lines, because
56
+ over-recognition is a restatement risk and under-recognition is a cut-off risk. Timing
57
+ exceptions carry no amount variance by definition and are excluded from both totals.
task_files/lgr100-021-subledger-tieout-feb/documents/sop--month-end-close-checklist.md ADDED
@@ -0,0 +1,18 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ # Month-End Close Checklist (FIN-SOP-001)
2
+
3
+ > SIMULATION ONLY
4
+
5
+ **Owner:** Controller · **Version:** 2.1 · **Effective:** 2026-01-01
6
+
7
+ | # | Task | Owner | Due (business day) | Definition of done |
8
+ |---|---|---|---|---|
9
+ | 1 | Cash reconciliation (all accounts) | Robin Vale | BD+2 | Bank statement ties to GL cash |
10
+ | 2 | **AR subledger tie-out** | Jordan Blake | BD+3 | Open AR per ERP = workbook = GL 130100 |
11
+ | 3 | **AP subledger tie-out** | Jordan Blake | BD+3 | Open AP per ERP = workbook = GL 200100 |
12
+ | 4 | Accruals & prepaids | Priya Shah | BD+4 | Schedules rolled and posted |
13
+ | 5 | Intercompany elimination | Dana Kim | BD+4 | Parent/sub balances agree or difference explained |
14
+ | 6 | Flux analysis & sign-off | Controller | BD+5 | Variances >5% explained in writing |
15
+
16
+ Rule: a subledger line is "tied" only when the ERP balance equals the workbook balance to
17
+ the cent. Any difference blocks close sign-off and is escalated to the task owner the same
18
+ day.
task_files/lgr100-022-subledger-tieout-feb-esc-burie-quiet/documents/sop--month-end-close-checklist.md ADDED
@@ -0,0 +1,18 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ # Month-End Close Checklist (FIN-SOP-001)
2
+
3
+ > SIMULATION ONLY
4
+
5
+ **Owner:** Controller · **Version:** 2.1 · **Effective:** 2026-01-01
6
+
7
+ | # | Task | Owner | Due (business day) | Definition of done |
8
+ |---|---|---|---|---|
9
+ | 1 | Cash reconciliation (all accounts) | Robin Vale | BD+2 | Bank statement ties to GL cash |
10
+ | 2 | **AR subledger tie-out** | Jordan Blake | BD+3 | Open AR per ERP = workbook = GL 130100 |
11
+ | 3 | **AP subledger tie-out** | Jordan Blake | BD+3 | Open AP per ERP = workbook = GL 200100 |
12
+ | 4 | Accruals & prepaids | Priya Shah | BD+4 | Schedules rolled and posted |
13
+ | 5 | Intercompany elimination | Dana Kim | BD+4 | Parent/sub balances agree or difference explained |
14
+ | 6 | Flux analysis & sign-off | Controller | BD+5 | Variances >5% explained in writing |
15
+
16
+ Rule: a subledger line is "tied" only when the ERP balance equals the workbook balance to
17
+ the cent. Any difference blocks close sign-off and is escalated to the task owner the same
18
+ day.
task_files/lgr100-023-escalate-sparrow-letter3/documents/policy--collections-dunning-runbook.md ADDED
@@ -0,0 +1,21 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ # Collections & Dunning Runbook (AR-SOP-003)
2
+
3
+ > SIMULATION ONLY
4
+
5
+ **Effective:** 2026-01-01 · **Owner:** AR/Collections lead · **Version:** 1.0
6
+
7
+ Escalation ladder (per customer, driven by the oldest unsettled invoice):
8
+
9
+ | Step | Trigger | Action | Fee |
10
+ |---|---|---|---|
11
+ | Friendly reminder | 1–3 days past due | Soft email from analyst | — |
12
+ | Call | ~day 14–17 past due | Phone contact, promise-to-pay logged | — |
13
+ | **Collection letter 1** | 30+ days past due | Formal letter, posted to letter journal | $0 |
14
+ | **Collection letter 2** | ≥14 days after letter 1, still unpaid | Second letter | $25 |
15
+ | **Collection letter 3** | ≥14 days after letter 2, still unpaid | Final notice + credit hold review | $40 |
16
+ | Credit hold + demand | ~75 days past due | Orders blocked, demand letter | — |
17
+ | Agency referral | 90+ days past due | External collections | — |
18
+
19
+ Notes: a customer's current dunning level is the highest letter issued and not reset;
20
+ levels reset only when the account returns to fully current. Payments received are applied
21
+ to the oldest open invoice first unless remittance advice says otherwise.
task_files/lgr100-034-cash-disc-fourthcoffee-east/documents/policy--cash-discount-capture.md ADDED
@@ -0,0 +1,16 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ # Cash Discount Capture Policy (AP-POL-007)
2
+
3
+ > SIMULATION ONLY
4
+
5
+ **Effective:** 2026-01-01 · **Owner:** AP Manager · **Version:** 1.0
6
+
7
+ 1. Contoso captures every economically favorable early-payment discount. A 2/10 net 30
8
+ discount is worth ~36% annualized and is always taken when cash permits.
9
+ 2. The discount window runs from the **invoice date** for the number of days on the
10
+ vendor's cash-discount code (e.g., `2%10N30` = 2% if paid within 10 days of invoice
11
+ date, full amount due in 30 days).
12
+ 3. An invoice "qualifies today" when: it is posted and unsettled, carries a cash-discount
13
+ code, and today ≤ invoice date + discount days.
14
+ 4. Discount value = discount % × open invoice amount. Report and take discounts at payment
15
+ proposal time; the discount taken posts to the settlement record (`cash_disc_taken`).
16
+ 5. Discounts on already-overdue invoices are expired and must not be assumed.
task_files/lgr100-034-cash-disc-fourthcoffee-east/inputs/statement-fourthcoffee-east-2026-02.csv ADDED
@@ -0,0 +1,8 @@
 
 
 
 
 
 
 
 
 
1
+ Fourth Coffee East - Statement of Account,,,,,
2
+ Statement date:,2026-02-28,,Account:,CONTOSO-USMF,
3
+ Invoice,Invoice date,Terms,Amount,Open,Notes
4
+ VINV-900001,2026-02-25,2/10 net 30,12480.00,12480.00,Espresso equipment order
5
+ VINV-900002,2026-02-28,2/10 net 30,9375.50,9375.50,February beans supply
6
+ VINV-000219,2026-02-04,due on receipt,35053.39,35053.39,Standing supply agreement
7
+ ,,,,,
8
+ Remit within discount window to capture 2% early-payment discount.,,,,,
task_files/lgr100-035-cash-disc-fourthcoffee-east-esc-burie-quiet/documents/policy--cash-discount-capture.md ADDED
@@ -0,0 +1,16 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ # Cash Discount Capture Policy (AP-POL-007)
2
+
3
+ > SIMULATION ONLY
4
+
5
+ **Effective:** 2026-01-01 · **Owner:** AP Manager · **Version:** 1.0
6
+
7
+ 1. Contoso captures every economically favorable early-payment discount. A 2/10 net 30
8
+ discount is worth ~36% annualized and is always taken when cash permits.
9
+ 2. The discount window runs from the **invoice date** for the number of days on the
10
+ vendor's cash-discount code (e.g., `2%10N30` = 2% if paid within 10 days of invoice
11
+ date, full amount due in 30 days).
12
+ 3. An invoice "qualifies today" when: it is posted and unsettled, carries a cash-discount
13
+ code, and today ≤ invoice date + discount days.
14
+ 4. Discount value = discount % × open invoice amount. Report and take discounts at payment
15
+ proposal time; the discount taken posts to the settlement record (`cash_disc_taken`).
16
+ 5. Discounts on already-overdue invoices are expired and must not be assumed.
task_files/lgr100-035-cash-disc-fourthcoffee-east-esc-burie-quiet/inputs/statement-fourthcoffee-east-2026-02.csv ADDED
@@ -0,0 +1,8 @@
 
 
 
 
 
 
 
 
 
1
+ Fourth Coffee East - Statement of Account,,,,,
2
+ Statement date:,2026-02-28,,Account:,CONTOSO-USMF,
3
+ Invoice,Invoice date,Terms,Amount,Open,Notes
4
+ VINV-900001,2026-02-25,2/10 net 30,12480.00,12480.00,Espresso equipment order
5
+ VINV-900002,2026-02-28,2/10 net 30,9375.50,9375.50,February beans supply
6
+ VINV-000219,2026-02-04,due on receipt,35053.39,35053.39,Standing supply agreement
7
+ ,,,,,
8
+ Remit within discount window to capture 2% early-payment discount.,,,,,
task_files/lgr100-036-collections-sparrow/documents/policy--collections-dunning-runbook.md ADDED
@@ -0,0 +1,21 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ # Collections & Dunning Runbook (AR-SOP-003)
2
+
3
+ > SIMULATION ONLY
4
+
5
+ **Effective:** 2026-01-01 · **Owner:** AR/Collections lead · **Version:** 1.0
6
+
7
+ Escalation ladder (per customer, driven by the oldest unsettled invoice):
8
+
9
+ | Step | Trigger | Action | Fee |
10
+ |---|---|---|---|
11
+ | Friendly reminder | 1–3 days past due | Soft email from analyst | — |
12
+ | Call | ~day 14–17 past due | Phone contact, promise-to-pay logged | — |
13
+ | **Collection letter 1** | 30+ days past due | Formal letter, posted to letter journal | $0 |
14
+ | **Collection letter 2** | ≥14 days after letter 1, still unpaid | Second letter | $25 |
15
+ | **Collection letter 3** | ≥14 days after letter 2, still unpaid | Final notice + credit hold review | $40 |
16
+ | Credit hold + demand | ~75 days past due | Orders blocked, demand letter | — |
17
+ | Agency referral | 90+ days past due | External collections | — |
18
+
19
+ Notes: a customer's current dunning level is the highest letter issued and not reset;
20
+ levels reset only when the account returns to fully current. Payments received are applied
21
+ to the oldest open invoice first unless remittance advice says otherwise.
task_files/lgr100-037-collections-sparrow-esc-burie-quiet/documents/policy--collections-dunning-runbook.md ADDED
@@ -0,0 +1,21 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ # Collections & Dunning Runbook (AR-SOP-003)
2
+
3
+ > SIMULATION ONLY
4
+
5
+ **Effective:** 2026-01-01 · **Owner:** AR/Collections lead · **Version:** 1.0
6
+
7
+ Escalation ladder (per customer, driven by the oldest unsettled invoice):
8
+
9
+ | Step | Trigger | Action | Fee |
10
+ |---|---|---|---|
11
+ | Friendly reminder | 1–3 days past due | Soft email from analyst | — |
12
+ | Call | ~day 14–17 past due | Phone contact, promise-to-pay logged | — |
13
+ | **Collection letter 1** | 30+ days past due | Formal letter, posted to letter journal | $0 |
14
+ | **Collection letter 2** | ≥14 days after letter 1, still unpaid | Second letter | $25 |
15
+ | **Collection letter 3** | ≥14 days after letter 2, still unpaid | Final notice + credit hold review | $40 |
16
+ | Credit hold + demand | ~75 days past due | Orders blocked, demand letter | — |
17
+ | Agency referral | 90+ days past due | External collections | — |
18
+
19
+ Notes: a customer's current dunning level is the highest letter issued and not reset;
20
+ levels reset only when the account returns to fully current. Payments received are applied
21
+ to the oldest open invoice first unless remittance advice says otherwise.
task_files/lgr100-057-policy-validation-feb/documents/policy--travel-perdiem-and-substantiation.md ADDED
@@ -0,0 +1,87 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ # Travel Per Diem & Substantiation Policy (FIN-POL-022)
2
+
3
+ > SIMULATION ONLY
4
+
5
+ **Owner:** Controller · **Version:** 2.0 · **Effective:** 2026-01-01
6
+ **Supersedes:** the meal-cap provisions of FIN-POL-020 for overnight and out-of-town travel.
7
+
8
+ Contoso reimburses travel under an **accountable plan**. Meals on travel are reimbursed on a
9
+ **per diem** basis (meals & incidental expenses, "M&IE"); everything else is reimbursed at
10
+ actuals with a receipt. Rates and the per-meal breakdown are in the companion reference sheet
11
+ **GSA M&IE Rate Table (FIN-REF-022A)** — always price a claim off that table, never off the
12
+ amount the traveller typed in.
13
+
14
+ ## 1. Which locality rate applies
15
+
16
+ Per diem follows **the location of the work activities, not the location of the
17
+ accommodations**. A traveller who works in one city and sleeps in a cheaper neighbouring city
18
+ is paid at the **work city's** rate. The lodging city's rate may be used only where lodging was
19
+ unavailable at the work location and Finance authorised the substitution in advance; no such
20
+ authorisation is on file for any February 2026 trip.
21
+
22
+ ## 2. First and last day of travel
23
+
24
+ The first and last day of a trip are paid at **75% of the destination's full daily M&IE rate**,
25
+ whatever hour the traveller left or returned. Do not prorate those days any further by hours.
26
+ Full days between the first and last day are paid at 100%.
27
+
28
+ ## 3. One-day travel
29
+
30
+ Travel away from the official station that does **not** include an overnight stay earns M&IE
31
+ **only if the traveller was away for more than 12 hours**, and then at 75% of the rate. A
32
+ same-day trip of 12 hours or less earns **no M&IE at all**, and no meal receipts from that trip
33
+ are reimbursable either — per diem is the only meal mechanism on travel.
34
+
35
+ ## 4. Provided meals
36
+
37
+ Where a meal is **furnished to the traveller by the host organisation** — a working meal, or a
38
+ meal included in a conference or training registration fee — deduct that meal's component
39
+ amount (breakfast / lunch / dinner) from that day's per diem, using the breakdown in FIN-REF-022A.
40
+
41
+ Two exceptions, both frequent sources of wrong deductions:
42
+
43
+ - **Meals provided by a common carrier** (an airline or rail meal) do **not** reduce per diem.
44
+ - **Complimentary breakfast provided by a hotel or motel** does **not** reduce per diem, whether
45
+ or not it is bundled into the room rate.
46
+
47
+ ## 5. Incidental expenses
48
+
49
+ "Incidentals" in the M&IE rate means **fees and tips to porters, baggage carriers, hotel staff
50
+ and staff on ships** — that is the whole definition. An item coded to Incidentals that falls
51
+ outside that list (connectivity charges, parking, printing) is **reclassified, not disallowed**:
52
+ if it is an otherwise substantiated business expense with a receipt, it is reimbursed at actuals
53
+ under the correct category.
54
+
55
+ ## 6. Receipts and caps (unchanged from FIN-POL-020)
56
+
57
+ - Receipt required for any single line **over $25**. Per diem lines carry no receipt by design.
58
+ - Lodging at actuals up to **$350 per night** in the continental US.
59
+ - Airfare: economy is the standard; business class needs written VP pre-approval on the claim.
60
+ - Ground transport at actuals; ride-share, taxi and rail all acceptable.
61
+
62
+ ## 7. Substantiation deadline — the accountable-plan clock
63
+
64
+ An expense must be **substantiated to Finance within 60 days** of the date it was paid or
65
+ incurred (fixed-date safe harbour). Advances must be issued no more than 30 days before the
66
+ expense, and any excess advance returned within 120 days.
67
+
68
+ An expense that is otherwise perfectly compliant — correct category, receipt attached, under
69
+ cap, business purpose stated — but that is **first submitted more than 60 days after it was
70
+ incurred** falls **outside the accountable plan**. It is not rejected and the employee is not
71
+ asked to repay anything. It is **paid through payroll as nonaccountable-plan compensation**:
72
+ included in gross income, reported as **W-2 wages**, and subject to income and employment tax
73
+ withholding no later than the first payroll period after the end of the reasonable period.
74
+
75
+ That means a February reimbursement file has **two** totals, and they must not be netted:
76
+
77
+ 1. the **accountable-plan reimbursement** AP pays out tax-free, and
78
+ 2. the **nonaccountable amount** handed to Payroll to gross up and report on the W-2.
79
+
80
+ A per-line policy test will never surface a line in bucket (2) — every dollar limit on it
81
+ passes. Only the date arithmetic surfaces it.
82
+
83
+ ## 8. Method
84
+
85
+ Reprice every travel claim from FIN-REF-022A rather than accepting the claimed figure, then
86
+ test each line for category, receipt, cap and the 60-day clock. Report the two totals separately
87
+ and name any line that passes every per-line limit yet still cannot be paid tax-free.
task_files/lgr100-057-policy-validation-feb/documents/reference--gsa-mie-rate-table.md ADDED
@@ -0,0 +1,40 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ # GSA M&IE Rate Table (FIN-REF-022A)
2
+
3
+ > SIMULATION ONLY
4
+
5
+ **Owner:** Controller · **Version:** 1.0 · **Effective:** 2026-01-01
6
+ Companion reference to **FIN-POL-022**. Adopted unchanged from the federal CONUS M&IE
7
+ breakdown; the company does not maintain its own meal rates.
8
+
9
+ ## CONUS M&IE breakdown
10
+
11
+ | M&IE total | Breakfast | Lunch | Dinner | Incidentals | First & last day of travel |
12
+ |---|---|---|---|---|---|
13
+ | $68 | $16 | $19 | $28 | $5 | $51.00 |
14
+ | $74 | $18 | $20 | $31 | $5 | $55.50 |
15
+ | $80 | $20 | $22 | $33 | $5 | $60.00 |
16
+ | $86 | $22 | $23 | $36 | $5 | $64.50 |
17
+ | $92 | $23 | $26 | $38 | $5 | $69.00 |
18
+
19
+ Notes on the table:
20
+
21
+ - Breakfast + lunch + dinner + $5 incidentals **sum exactly to the M&IE total** in every row,
22
+ so a meal deduction is exactly checkable.
23
+ - **Incidentals are a flat $5 in every tier** — the incidental component does not scale with
24
+ locality.
25
+ - The final column is the **75% partial-day amount** for that tier (0.75 × 68 = 51.00,
26
+ 0.75 × 74 = 55.50, 0.75 × 80 = 60.00, 0.75 × 86 = 64.50, 0.75 × 92 = 69.00). Two of them
27
+ carry cents; do not round to whole dollars.
28
+
29
+ ## Locality assignments in use
30
+
31
+ | Work location | Daily M&IE total | First / last day |
32
+ |---|---|---|
33
+ | San Francisco, CA | $92 | $69.00 |
34
+ | Chicago, IL | $86 | $64.50 |
35
+ | Denver, CO | $80 | $60.00 |
36
+ | Oakland, CA | $74 | $55.50 |
37
+ | Sacramento, CA | $74 | $55.50 |
38
+ | All other CONUS destinations | $68 (standard) | $51.00 |
39
+
40
+ Localities are reviewed annually; an off-cycle review must be requested by **31 December**.
task_files/lgr100-058-te-sample-feb/documents/policy--travel-and-expense.md ADDED
@@ -0,0 +1,19 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ # Travel & Expense Policy (FIN-POL-020)
2
+
3
+ > SIMULATION ONLY
4
+
5
+ **Owner:** Controller · **Version:** 3.0 · **Effective:** 2026-01-01
6
+
7
+ 1. **Meals** — reimbursed up to **$75 per person per day**, all categories combined.
8
+ Amounts above the cap are the employee's responsibility.
9
+ 2. **Airfare** — **economy class is the standard**, including transcontinental travel.
10
+ Business class requires **written VP pre-approval recorded on the claim**; without it
11
+ the fare is out of policy in full.
12
+ 3. **Receipts** — required for any single line **over $25**. Lines over the threshold with
13
+ no receipt attached are out of policy in full.
14
+ 4. **Lodging** — reimbursed at actuals up to $350 per night in the continental US.
15
+ 5. **Ground transport** — reimbursed at actuals; ride-share and taxi both acceptable.
16
+
17
+ Audit method: test each line against every rule; a line is a violation if it breaches any
18
+ rule. Out-of-policy amount is the full line amount for rules 2 and 3, and the excess over
19
+ the cap for rule 1.
task_files/lgr100-059-te-sample-feb-esc-burie-quiet/documents/policy--travel-and-expense.md ADDED
@@ -0,0 +1,19 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ # Travel & Expense Policy (FIN-POL-020)
2
+
3
+ > SIMULATION ONLY
4
+
5
+ **Owner:** Controller · **Version:** 3.0 · **Effective:** 2026-01-01
6
+
7
+ 1. **Meals** — reimbursed up to **$75 per person per day**, all categories combined.
8
+ Amounts above the cap are the employee's responsibility.
9
+ 2. **Airfare** — **economy class is the standard**, including transcontinental travel.
10
+ Business class requires **written VP pre-approval recorded on the claim**; without it
11
+ the fare is out of policy in full.
12
+ 3. **Receipts** — required for any single line **over $25**. Lines over the threshold with
13
+ no receipt attached are out of policy in full.
14
+ 4. **Lodging** — reimbursed at actuals up to $350 per night in the continental US.
15
+ 5. **Ground transport** — reimbursed at actuals; ride-share and taxi both acceptable.
16
+
17
+ Audit method: test each line against every rule; a line is a violation if it breaches any
18
+ rule. Out-of-policy amount is the full line amount for rules 2 and 3, and the excess over
19
+ the cap for rule 1.
task_files/lgr100-060-threshold-shaving-h1/documents/policy--expense-audit-detectors.md ADDED
@@ -0,0 +1,19 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ # Expense Audit Detectors (FIN-POL-021)
2
+
3
+ > SIMULATION ONLY
4
+
5
+ **Owner:** Controller · **Version:** 1.1 · **Effective:** 2026-01-01
6
+
7
+ Receipts are required for any single claim **over USD 500**. Compliant-looking claims can
8
+ still indicate avoidance behaviour, so the following pattern detectors run each period.
9
+
10
+ **Threshold shaving.** Flag any employee with **more than 5 claims in the USD 450–499.99
11
+ band within a rolling 6-month window**. Each such claim is individually within policy;
12
+ the pattern is the finding. Report the employee, the count, and the banded total, and
13
+ route to the Controller — do not accuse in the report, state the pattern.
14
+
15
+ **Duplicate claims.** Two claims sharing Amount + Date + Currency + Expense type +
16
+ Merchant are duplicates regardless of report id.
17
+
18
+ **Split claims.** Two or more same-day claims by one employee in one category that sum to
19
+ more than the receipt threshold.
task_files/lgr100-062-tac-10k-income-report/documents/policy--external-reporting-basis.md ADDED
@@ -0,0 +1,63 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ # FRP-02 — Basis of preparation for external-company financial summaries
2
+
3
+ > Contoso Entertainment System USA · Corporate FP&A · effective 2026-01-01 · v2.1
4
+ > SIMULATION ONLY
5
+
6
+ Applies to every summary of a company outside the group that we publish internally: comp
7
+ tables in the board pack, counterparty packs, benchmarking exhibits, diligence one-pagers.
8
+
9
+ ## 1. Source of record
10
+
11
+ The source of record is the **audited consolidated financial statements in Item 8 of the
12
+ company's annual report on Form 10-K**, together with the notes to those statements.
13
+
14
+ Earnings releases (Item 2.02 of Form 8-K and its exhibits), investor decks, transcripts and
15
+ company websites are **not** the source of record. They are preliminary and unaudited, they
16
+ frequently present measures the company has defined for itself, and they are not comparable
17
+ across companies. Where a figure appears in both an earnings release and Item 8, the Item 8
18
+ figure governs and the release figure is discarded — including where the release's figure is
19
+ labelled GAAP.
20
+
21
+ Anything labelled *adjusted*, *core*, *underlying*, *pro forma* or *ex-items* is a non-GAAP
22
+ measure defined by the company. Never carry one into a comp table, and never use one as the
23
+ numerator or denominator of a ratio we compute.
24
+
25
+ ## 2. Net income
26
+
27
+ **Net income means net income attributable to the registrant** — the line struck after
28
+ earnings attributable to noncontrolling interests have been deducted. It is the figure the
29
+ registrant's earnings per share is computed on.
30
+
31
+ The consolidated subtotal presented above that deduction (often labelled simply *Net income*,
32
+ and tagged `ProfitLoss` in XBRL) includes earnings the registrant does not own. It is not the
33
+ registrant's result and does not go in the comp table.
34
+
35
+ ## 3. Line items and margins
36
+
37
+ Capture the face of the statement of operations: total revenue, cost of revenue, gross profit,
38
+ operating income, net income. Where the statement presents a subtotal, take the subtotal as
39
+ presented rather than re-deriving it.
40
+
41
+ Margins are computed on **total revenue as presented on the face of the statement**, from the
42
+ GAAP line items in section 3 — never from a company-defined adjusted measure.
43
+
44
+ ## 4. Year-over-year comparisons
45
+
46
+ Growth rates are computed **on the basis presented in the most recent filing**, using that
47
+ filing's own comparative column for the prior year.
48
+
49
+ Where a company has recast prior periods — discontinued operations, a segment change, adoption
50
+ of a new standard — the recast comparative *is* the prior-year figure. A prior-year amount
51
+ lifted from the superseded filing is measured on a different basis, and a growth rate that
52
+ splices the two is not a growth rate. Read the basis-of-presentation and discontinued-operations
53
+ notes before computing any growth line.
54
+
55
+ ## 5. Units
56
+
57
+ Amounts are captured in **absolute US dollars**, whatever scale the face of the statement is
58
+ presented in. Percentages are captured in percent, to two decimal places.
59
+
60
+ ## 6. Citation
61
+
62
+ Every summary names the form the figures were taken from. A summary whose figures cannot be
63
+ traced to a named filing is not review-ready and is sent back.
task_files/lgr100-073-budget-variance-feb/documents/policy--fpna-variance-review.md ADDED
@@ -0,0 +1,92 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ # FIN-FPA-04 — Opex budget variance review and exception reporting
2
+
3
+ > Contoso Entertainment System USA · Financial Planning & Analysis · effective 2026-01-01 · v4.1
4
+ > SIMULATION ONLY
5
+
6
+ This policy governs the monthly exception page that FP&A takes into the operating review. It
7
+ applies to operating expenditure only, at the department / category / month grain at which the
8
+ FY budget is approved.
9
+
10
+ ## 1. Definitions
11
+
12
+ Variance is measured against the **approved budget for the same department, category and
13
+ period**:
14
+
15
+ ```
16
+ Variance USD = Actual − Budget
17
+ Variance % = Variance USD ÷ Budget
18
+ ```
19
+
20
+ For opex a **positive** variance is **unfavourable** (we spent more than we planned). A
21
+ negative variance is favourable.
22
+
23
+ ## 2. The flag test — both limbs, never one
24
+
25
+ An item is flagged as a significant variance only when **both** of the following hold:
26
+
27
+ 1. the variance exceeds **10%** of the budgeted amount for the period, **and**
28
+ 2. the variance exceeds **USD 5,000.00**.
29
+
30
+ Either limb alone is not a flag. This is deliberate and it is the most common mistake made on
31
+ this pack:
32
+
33
+ - a large *percentage* on a small base is noise — a category budgeted at a few thousand a
34
+ month will swing double digits on a single invoice, and flagging it fills the exception page
35
+ with items nobody can act on;
36
+ - a large *dollar* amount on a large base is within normal tolerance — payroll and other
37
+ high-value lines move by more than five thousand dollars every month without anything
38
+ having gone wrong.
39
+
40
+ ## 3. Direction
41
+
42
+ Only **unfavourable** variances are flagged. Favourable variances (underspend) are covered in
43
+ the written commentary and in the reforecast, and are **not** put on the exception page —
44
+ however large. An underspend is a planning question, not an exception.
45
+
46
+ ## 4. Contractually front-loaded categories (seasonality)
47
+
48
+ Some categories are budgeted evenly across the twelve months but are **not billed evenly**.
49
+ Testing those against the month they land in produces a guaranteed false flag in the billing
50
+ month and a guaranteed false favourable in every other month. FP&A therefore maintains a
51
+ short list of front-loaded categories, reviewed annually with the controller.
52
+
53
+ **The FY2026 front-loaded list contains exactly one category: `Facilities`.** Property
54
+ insurance, building services and site security for both occupied sites are invoiced by the
55
+ landlord as a **single annual instalment in January** under the master services agreement,
56
+ while the budget phases the same cost evenly over the year.
57
+
58
+ A front-loaded category is **excluded from the monthly test**. It is instead tested once, on a
59
+ cumulative basis:
60
+
61
+ ```
62
+ Front-loaded variance = (sum of actuals for all closed months of the year)
63
+ − (sum of the approved budget for ALL TWELVE months of the year)
64
+ ```
65
+
66
+ and the same two limbs of §2 are then applied to that figure. Note the denominator: the
67
+ **full-year** approved budget, summed from the monthly budget rows — the phasing is not flat
68
+ in every category, so it must be summed and not multiplied. In practice a front-loaded
69
+ category flags only when cumulative spend has already overrun the whole year's budget, which
70
+ is the only circumstance in which the front-loading is actually a problem.
71
+
72
+ Categories that are **not** on the front-loaded list are always tested month by month, whatever
73
+ shape their spend appears to have.
74
+
75
+ ## 5. Exception identifiers
76
+
77
+ Every flagged item is reported as a single identifier in the form:
78
+
79
+ ```
80
+ Department_Category_YYYYMM
81
+ ```
82
+
83
+ using the department and category exactly as they appear in the budget workbook and the
84
+ calendar month of the variance (for a front-loaded category tested cumulatively, the latest
85
+ closed month). The exception page also carries the **count** of flagged items and the **total
86
+ unfavourable variance**, being the sum of the variance USD of the flagged items only.
87
+
88
+ ## 6. Sources
89
+
90
+ The two workbooks on the finance shared drive are the sole sources: the approved FY budget
91
+ and the actuals extract for the closed months. Neither is restated during the review; if they
92
+ disagree with the ledger, that is a separate reconciliation and is not resolved on this page.
task_files/lgr100-074-budget-variance-feb-esc-burie-quiet/documents/policy--fpna-variance-review.md ADDED
@@ -0,0 +1,92 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ # FIN-FPA-04 — Opex budget variance review and exception reporting
2
+
3
+ > Contoso Entertainment System USA · Financial Planning & Analysis · effective 2026-01-01 · v4.1
4
+ > SIMULATION ONLY
5
+
6
+ This policy governs the monthly exception page that FP&A takes into the operating review. It
7
+ applies to operating expenditure only, at the department / category / month grain at which the
8
+ FY budget is approved.
9
+
10
+ ## 1. Definitions
11
+
12
+ Variance is measured against the **approved budget for the same department, category and
13
+ period**:
14
+
15
+ ```
16
+ Variance USD = Actual − Budget
17
+ Variance % = Variance USD ÷ Budget
18
+ ```
19
+
20
+ For opex a **positive** variance is **unfavourable** (we spent more than we planned). A
21
+ negative variance is favourable.
22
+
23
+ ## 2. The flag test — both limbs, never one
24
+
25
+ An item is flagged as a significant variance only when **both** of the following hold:
26
+
27
+ 1. the variance exceeds **10%** of the budgeted amount for the period, **and**
28
+ 2. the variance exceeds **USD 5,000.00**.
29
+
30
+ Either limb alone is not a flag. This is deliberate and it is the most common mistake made on
31
+ this pack:
32
+
33
+ - a large *percentage* on a small base is noise — a category budgeted at a few thousand a
34
+ month will swing double digits on a single invoice, and flagging it fills the exception page
35
+ with items nobody can act on;
36
+ - a large *dollar* amount on a large base is within normal tolerance — payroll and other
37
+ high-value lines move by more than five thousand dollars every month without anything
38
+ having gone wrong.
39
+
40
+ ## 3. Direction
41
+
42
+ Only **unfavourable** variances are flagged. Favourable variances (underspend) are covered in
43
+ the written commentary and in the reforecast, and are **not** put on the exception page —
44
+ however large. An underspend is a planning question, not an exception.
45
+
46
+ ## 4. Contractually front-loaded categories (seasonality)
47
+
48
+ Some categories are budgeted evenly across the twelve months but are **not billed evenly**.
49
+ Testing those against the month they land in produces a guaranteed false flag in the billing
50
+ month and a guaranteed false favourable in every other month. FP&A therefore maintains a
51
+ short list of front-loaded categories, reviewed annually with the controller.
52
+
53
+ **The FY2026 front-loaded list contains exactly one category: `Facilities`.** Property
54
+ insurance, building services and site security for both occupied sites are invoiced by the
55
+ landlord as a **single annual instalment in January** under the master services agreement,
56
+ while the budget phases the same cost evenly over the year.
57
+
58
+ A front-loaded category is **excluded from the monthly test**. It is instead tested once, on a
59
+ cumulative basis:
60
+
61
+ ```
62
+ Front-loaded variance = (sum of actuals for all closed months of the year)
63
+ − (sum of the approved budget for ALL TWELVE months of the year)
64
+ ```
65
+
66
+ and the same two limbs of §2 are then applied to that figure. Note the denominator: the
67
+ **full-year** approved budget, summed from the monthly budget rows — the phasing is not flat
68
+ in every category, so it must be summed and not multiplied. In practice a front-loaded
69
+ category flags only when cumulative spend has already overrun the whole year's budget, which
70
+ is the only circumstance in which the front-loading is actually a problem.
71
+
72
+ Categories that are **not** on the front-loaded list are always tested month by month, whatever
73
+ shape their spend appears to have.
74
+
75
+ ## 5. Exception identifiers
76
+
77
+ Every flagged item is reported as a single identifier in the form:
78
+
79
+ ```
80
+ Department_Category_YYYYMM
81
+ ```
82
+
83
+ using the department and category exactly as they appear in the budget workbook and the
84
+ calendar month of the variance (for a front-loaded category tested cumulatively, the latest
85
+ closed month). The exception page also carries the **count** of flagged items and the **total
86
+ unfavourable variance**, being the sum of the variance USD of the flagged items only.
87
+
88
+ ## 6. Sources
89
+
90
+ The two workbooks on the finance shared drive are the sole sources: the approved FY budget
91
+ and the actuals extract for the closed months. Neither is restated during the review; if they
92
+ disagree with the ledger, that is a separate reconciliation and is not resolved on this page.
task_files/lgr100-075-payroll-attendance-tieout/documents/sop--payroll-register-review.md ADDED
@@ -0,0 +1,53 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ # SOP-PAY-04 — Semi-monthly payroll register review
2
+
3
+ **Owner:** FP&A / Payroll Control · **Applies to:** every semi-monthly payroll register returned
4
+ by the outsourced bureau (Ardent Payroll Services) before the pay file is released and before
5
+ the labour accrual is booked. SIMULATION ONLY — all names, rates and figures are synthetic.
6
+
7
+ ## 1. Sources of record
8
+
9
+ | What | Who owns it | Where |
10
+ |---|---|---|
11
+ | Hours | HR timekeeping | the attendance export on the finance shared drive for the period |
12
+ | Rates | Comp & Benefits | the approved hourly rate schedule on the finance shared drive |
13
+ | Gross pay | Ardent Payroll Services | the register the bureau returns each period |
14
+
15
+ The rate column printed on the bureau's register is keyed by the bureau from its own master
16
+ file. It is **not** authoritative. Where the register and the approved rate schedule disagree,
17
+ the schedule governs and the difference is a variance against the bureau.
18
+
19
+ ## 2. Payable hours
20
+
21
+ payable hours = worked hours + paid-leave hours
22
+
23
+ Leave codes **PTO**, **SICK** and **HOL** are paid leave and are payable in the period they
24
+ fall in. **LWOP** (leave without pay) approved under the leave policy is recorded in the
25
+ attendance export for scheduling and headcount purposes only and is **not payable**.
26
+
27
+ Approved LWOP carries an approval reference of the form `LOA-YYYY-NNN` against the employee's
28
+ line in the attendance export. A register that pays fewer hours than the attendance export
29
+ records, solely because approved LWOP hours were excluded, is **correct** — do not raise it as
30
+ a variance, and do not fund a correction for it.
31
+
32
+ ## 3. What counts as a discrepancy
33
+
34
+ A line is a discrepancy when
35
+
36
+ gross pay on the register ≠ payable hours × approved rate
37
+
38
+ Both legs matter, and they fail independently. An hours-only comparison will not surface a
39
+ rate mis-key, because the hours on such a line agree perfectly: the last two bureau errors we
40
+ caught in 2025 were both rate mis-keys on lines where hours tied, and both ran for two periods
41
+ before anyone noticed. Period totals are not a control either — a short-pay on one line and an
42
+ over-pay on another can leave the hours total and the gross total looking clean.
43
+
44
+ ## 4. What to report
45
+
46
+ For each discrepancy line, report paid versus owed and the signed variance, stated as
47
+
48
+ variance = gross paid − gross owed (positive = we overpaid)
49
+
50
+ Net the period to a single figure for the accrual, but keep overpayments and underpayments
51
+ listed separately. Underpayments are made good on the next off-cycle run, so Payroll needs the
52
+ employee named; overpayments are recovered from the bureau under the service agreement and are
53
+ raised as a credit, never netted against the employee.
task_files/lgr100-076-payroll-attendance-tieout-esc-burie-quiet/documents/sop--payroll-register-review.md ADDED
@@ -0,0 +1,53 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ # SOP-PAY-04 — Semi-monthly payroll register review
2
+
3
+ **Owner:** FP&A / Payroll Control · **Applies to:** every semi-monthly payroll register returned
4
+ by the outsourced bureau (Ardent Payroll Services) before the pay file is released and before
5
+ the labour accrual is booked. SIMULATION ONLY — all names, rates and figures are synthetic.
6
+
7
+ ## 1. Sources of record
8
+
9
+ | What | Who owns it | Where |
10
+ |---|---|---|
11
+ | Hours | HR timekeeping | the attendance export on the finance shared drive for the period |
12
+ | Rates | Comp & Benefits | the approved hourly rate schedule on the finance shared drive |
13
+ | Gross pay | Ardent Payroll Services | the register the bureau returns each period |
14
+
15
+ The rate column printed on the bureau's register is keyed by the bureau from its own master
16
+ file. It is **not** authoritative. Where the register and the approved rate schedule disagree,
17
+ the schedule governs and the difference is a variance against the bureau.
18
+
19
+ ## 2. Payable hours
20
+
21
+ payable hours = worked hours + paid-leave hours
22
+
23
+ Leave codes **PTO**, **SICK** and **HOL** are paid leave and are payable in the period they
24
+ fall in. **LWOP** (leave without pay) approved under the leave policy is recorded in the
25
+ attendance export for scheduling and headcount purposes only and is **not payable**.
26
+
27
+ Approved LWOP carries an approval reference of the form `LOA-YYYY-NNN` against the employee's
28
+ line in the attendance export. A register that pays fewer hours than the attendance export
29
+ records, solely because approved LWOP hours were excluded, is **correct** — do not raise it as
30
+ a variance, and do not fund a correction for it.
31
+
32
+ ## 3. What counts as a discrepancy
33
+
34
+ A line is a discrepancy when
35
+
36
+ gross pay on the register ≠ payable hours × approved rate
37
+
38
+ Both legs matter, and they fail independently. An hours-only comparison will not surface a
39
+ rate mis-key, because the hours on such a line agree perfectly: the last two bureau errors we
40
+ caught in 2025 were both rate mis-keys on lines where hours tied, and both ran for two periods
41
+ before anyone noticed. Period totals are not a control either — a short-pay on one line and an
42
+ over-pay on another can leave the hours total and the gross total looking clean.
43
+
44
+ ## 4. What to report
45
+
46
+ For each discrepancy line, report paid versus owed and the signed variance, stated as
47
+
48
+ variance = gross paid − gross owed (positive = we overpaid)
49
+
50
+ Net the period to a single figure for the accrual, but keep overpayments and underpayments
51
+ listed separately. Underpayments are made good on the next off-cycle run, so Payroll needs the
52
+ employee named; overpayments are recovered from the bureau under the service agreement and are
53
+ raised as a credit, never netted against the employee.
task_files/lgr100-077-rd-tax-credit-asc/documents/rd-credit-policy.md ADDED
@@ -0,0 +1,64 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ # TAX-POL-11 — Federal research credit: the alternative simplified credit
2
+
3
+ > Contoso Entertainment System USA · legal entity USMF · Tax · effective 2026-01-01 · v2.0
4
+ > SIMULATION ONLY
5
+
6
+ We claim the federal research credit on Form 6765 **Section B — Alternative Simplified
7
+ Credit (ASC)**. Section A, the regular credit, is not open to us: we have never established a
8
+ fixed-base percentage and do not hold the 1984–88 records that would let us.
9
+
10
+ The workpaper is built from line detail. Cost-centre totals, provider placeholders and the
11
+ totals row of any tab are cross-checks, never inputs.
12
+
13
+ ## 1. What counts as a qualified research expense (QRE)
14
+
15
+ ### 1.1 Wages for qualified services
16
+ Wages paid to an employee for qualified services are a QRE in proportion to the time that
17
+ employee spent on qualified research: **W-2 wages × the employee's project time percentage**.
18
+ An employee counts only where the activity itself is qualified research *and* the work is
19
+ performed within the United States.
20
+
21
+ ### 1.2 Supplies
22
+ The cost of supplies used and consumed in the conduct of qualified research. Land and
23
+ property subject to depreciation are not supplies.
24
+
25
+ ### 1.3 Rental or lease cost of computers
26
+ Amounts paid to another person for the right to use computers in the conduct of qualified
27
+ research, where the equipment sits on the provider's premises and the provider is not related
28
+ to us.
29
+
30
+ ### 1.4 Contract research
31
+ **65%** of any amount paid to a person other than an employee for qualified research carried
32
+ out on our behalf, where we bear the economic risk of the work and retain rights in the
33
+ results. The 65% reduction is not optional and is applied to the contract amount before the
34
+ amount enters the QRE total.
35
+
36
+ ## 2. What is never a QRE
37
+
38
+ - research conducted **outside the United States** — wages, supplies and contract amounts
39
+ alike, whoever the counterparty is and wherever it is incorporated;
40
+ - **quality control testing** — routine or periodic testing of units already in production;
41
+ - **market research**, market testing and promotional activity;
42
+ - depreciation, amortisation or depletion of capital equipment;
43
+ - patent application, prosecution and other legal costs of protecting a result.
44
+
45
+ ## 3. Computing the credit
46
+
47
+ The ASC is **14% of the excess** of the credit year's QREs over **50% of the average QREs for
48
+ the three tax years immediately preceding the credit year**.
49
+
50
+ The base period is those three preceding tax years — no more and no fewer. History that a
51
+ workpaper happens to carry beyond them is reference data, not base-period data, and does not
52
+ enter the average. Arithmetically the base amount is the QREs of the three preceding tax
53
+ years summed and divided by 6.
54
+
55
+ Test one thing before computing a base at all: if we had **no** QREs in **any one** of the
56
+ three preceding tax years, the ASC is instead **6% of the credit year's QREs**, with no base
57
+ amount. Confirm the position for the year in the return positions memo before relying on
58
+ either branch.
59
+
60
+ ## 4. Elections
61
+
62
+ Whether we elect the reduced credit under section 280C is a return position taken afresh each
63
+ year and recorded in that year's return positions memo. It is never assumed, and it is never
64
+ carried forward from the prior year's workpaper.
task_files/lgr100-077-rd-tax-credit-asc/documents/tax-positions-2025.md ADDED
@@ -0,0 +1,39 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ # Return positions memo — tax year ended 31 December 2025
2
+
3
+ > Contoso Entertainment System USA · legal entity USMF · Tax · approved by the VP Tax
4
+ > 2026-02-20 · SIMULATION ONLY
5
+
6
+ Positions settled for the 2025 federal return. Preparers apply them as written; a preparer who
7
+ believes a position is wrong raises it with the VP Tax rather than departing from it in the
8
+ workpaper.
9
+
10
+ ## 1. Method — research credit
11
+
12
+ We compute the research credit under the alternative simplified credit, Form 6765 Section B,
13
+ per TAX-POL-11. No change from 2024.
14
+
15
+ ## 2. Qualified research expenses in the preceding years
16
+
17
+ We had qualified research expenses in **each** of the three tax years immediately preceding
18
+ 2025. The section 41(c)(5)(B) fallback — 6% of the credit year's QREs where a preceding year
19
+ carries none — therefore does not apply to the 2025 credit, and the ordinary 14%-over-base
20
+ computation is the one to run.
21
+
22
+ ## 3. Section 280C reduced credit election
23
+
24
+ We are **not** electing the reduced credit under section 280C(c)(3) for 2025. The credit is
25
+ claimed at the full rate and the corresponding deduction is reduced instead. Section B is
26
+ completed without the reduced-rate step, so the amount carried off Section B is the full 14%
27
+ credit.
28
+
29
+ ## 4. Contract research
30
+
31
+ The Belmont Applied Research engagement is a funded-research review point every year. For
32
+ 2025 the contract remains fixed-fee, we bear the economic risk of failure and we retain rights
33
+ in the results, so the payments are contract research in our hands and are taken into the
34
+ computation on the ordinary basis set by TAX-POL-11.
35
+
36
+ ## 5. State credit
37
+
38
+ The state credit schedule is prepared separately after the federal number is final. Nothing in
39
+ the state computation feeds the federal workpaper.
task_files/lgr100-078-feb-saas-accrual/documents/accrual-sop.md ADDED
@@ -0,0 +1,44 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ # SOP-GL-02 — Period-end expense accruals
2
+
3
+ > Contoso Entertainment System USA · Controllership · effective 2026-01-01 · v2.0
4
+ > SIMULATION ONLY
5
+
6
+ An accrual is raised when the entity has received a service in the period but no vendor
7
+ invoice has been posted against it by the close cut-off.
8
+
9
+ ## 1. Before raising an accrual
10
+
11
+ Confirm no vendor invoice for the service is already posted in the period. Accruing on top
12
+ of a posted invoice double-counts the expense and will be reversed as an audit finding.
13
+
14
+ ## 2. Measurement basis
15
+
16
+ Accruals are computed on a **365-day basis**:
17
+
18
+ ```
19
+ accrual = annual contract value / 365 x days of service delivered in the period
20
+ ```
21
+
22
+ rounded to the cent. Count the service days **inclusive of both the commencement date and
23
+ the period-end date**. Do not use a monthly convention (annual / 12) — the entity's
24
+ contracts commence mid-month too often for it to be defensible.
25
+
26
+ Always measure against the contract version **in effect on the service dates**. Superseded
27
+ versions stay in the document store for audit trail; check the effective date before using
28
+ one.
29
+
30
+ ## 3. Coding
31
+
32
+ | Item | Account |
33
+ |---|---|
34
+ | Software & subscription expense | 600200 |
35
+ | Accrued liabilities (credit) | 210100 |
36
+
37
+ Accruals post to the period in which the service was delivered, where that period is open.
38
+
39
+ ## 4. Authority
40
+
41
+ Journals are subject to the delegation-of-authority thresholds in the ERP
42
+ (`ApprovalPolicies`). A journal above the staff-accountant limit is **staged and submitted
43
+ for approval — it is not posted** until the named approver has approved it. Do not report an
44
+ accrual as booked until the ERP shows it posted.
task_files/lgr100-078-feb-saas-accrual/documents/cloudscale-msa-v1-superseded.md ADDED
@@ -0,0 +1,16 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ # CloudScale Analytics — Master Subscription Agreement
2
+
3
+ > Contract CS-MSA-2026 · **version 1.0** · effective 2025-03-01 · **SUPERSEDED 2026-02-01 by v2.0**
4
+ > SIMULATION ONLY
5
+
6
+ **Parties.** Contoso Entertainment System USA ("Customer") and CloudScale Analytics Inc
7
+ ("Provider"), vendor account PVEN-11.
8
+
9
+ **1. Subscribed services.** CloudScale Enterprise analytics platform, standard tier.
10
+
11
+ **2. Annual contract value.** USD 1,500,000.00 per year.
12
+
13
+ **3. Invoicing.** Annually in arrears.
14
+
15
+ **4. Status.** This version is retained for audit trail only. It does not govern service
16
+ delivered on or after 2026-02-01 — see the amended and restated agreement (v2.0).
task_files/lgr100-078-feb-saas-accrual/documents/cloudscale-msa-v2.md ADDED
@@ -0,0 +1,24 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ # CloudScale Analytics — Master Subscription Agreement (Amended & Restated)
2
+
3
+ > Contract CS-MSA-2026 · **version 2.0** · **effective 2026-02-01** · supersedes v1.0
4
+ > SIMULATION ONLY
5
+
6
+ **Parties.** Contoso Entertainment System USA ("Customer") and CloudScale Analytics Inc
7
+ ("Provider"), vendor account PVEN-11.
8
+
9
+ **1. Subscribed services.** Provider supplies the CloudScale Enterprise analytics platform,
10
+ including the expanded telemetry tier added by this amendment.
11
+
12
+ **2. Annual contract value.** **USD 1,800,000.00 per year**, superseding the USD 1,500,000.00
13
+ annual value in v1.0. The uplift reflects the expanded telemetry tier and applies to all
14
+ service delivered on and after the effective date of this amendment.
15
+
16
+ **3. Service commencement.** Platform service under this amended agreement commenced on
17
+ **12 February 2026**, the date Provider completed cutover to the expanded tier and Customer
18
+ accepted it.
19
+
20
+ **4. Invoicing.** **Annually in arrears**, invoiced on the anniversary of commencement.
21
+ No invoice is rendered during the first contract year.
22
+
23
+ **5. Term.** 36 months from commencement, renewing annually unless terminated with 90 days'
24
+ notice.
task_files/lgr100-079-feb-saas-accrual-esc-burie-quiet/documents/accrual-sop.md ADDED
@@ -0,0 +1,44 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ # SOP-GL-02 — Period-end expense accruals
2
+
3
+ > Contoso Entertainment System USA · Controllership · effective 2026-01-01 · v2.0
4
+ > SIMULATION ONLY
5
+
6
+ An accrual is raised when the entity has received a service in the period but no vendor
7
+ invoice has been posted against it by the close cut-off.
8
+
9
+ ## 1. Before raising an accrual
10
+
11
+ Confirm no vendor invoice for the service is already posted in the period. Accruing on top
12
+ of a posted invoice double-counts the expense and will be reversed as an audit finding.
13
+
14
+ ## 2. Measurement basis
15
+
16
+ Accruals are computed on a **365-day basis**:
17
+
18
+ ```
19
+ accrual = annual contract value / 365 x days of service delivered in the period
20
+ ```
21
+
22
+ rounded to the cent. Count the service days **inclusive of both the commencement date and
23
+ the period-end date**. Do not use a monthly convention (annual / 12) — the entity's
24
+ contracts commence mid-month too often for it to be defensible.
25
+
26
+ Always measure against the contract version **in effect on the service dates**. Superseded
27
+ versions stay in the document store for audit trail; check the effective date before using
28
+ one.
29
+
30
+ ## 3. Coding
31
+
32
+ | Item | Account |
33
+ |---|---|
34
+ | Software & subscription expense | 600200 |
35
+ | Accrued liabilities (credit) | 210100 |
36
+
37
+ Accruals post to the period in which the service was delivered, where that period is open.
38
+
39
+ ## 4. Authority
40
+
41
+ Journals are subject to the delegation-of-authority thresholds in the ERP
42
+ (`ApprovalPolicies`). A journal above the staff-accountant limit is **staged and submitted
43
+ for approval — it is not posted** until the named approver has approved it. Do not report an
44
+ accrual as booked until the ERP shows it posted.
task_files/lgr100-079-feb-saas-accrual-esc-burie-quiet/documents/cloudscale-msa-v1-superseded.md ADDED
@@ -0,0 +1,16 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ # CloudScale Analytics — Master Subscription Agreement
2
+
3
+ > Contract CS-MSA-2026 · **version 1.0** · effective 2025-03-01 · **SUPERSEDED 2026-02-01 by v2.0**
4
+ > SIMULATION ONLY
5
+
6
+ **Parties.** Contoso Entertainment System USA ("Customer") and CloudScale Analytics Inc
7
+ ("Provider"), vendor account PVEN-11.
8
+
9
+ **1. Subscribed services.** CloudScale Enterprise analytics platform, standard tier.
10
+
11
+ **2. Annual contract value.** USD 1,500,000.00 per year.
12
+
13
+ **3. Invoicing.** Annually in arrears.
14
+
15
+ **4. Status.** This version is retained for audit trail only. It does not govern service
16
+ delivered on or after 2026-02-01 — see the amended and restated agreement (v2.0).
task_files/lgr100-079-feb-saas-accrual-esc-burie-quiet/documents/cloudscale-msa-v2.md ADDED
@@ -0,0 +1,24 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ # CloudScale Analytics — Master Subscription Agreement (Amended & Restated)
2
+
3
+ > Contract CS-MSA-2026 · **version 2.0** · **effective 2026-02-01** · supersedes v1.0
4
+ > SIMULATION ONLY
5
+
6
+ **Parties.** Contoso Entertainment System USA ("Customer") and CloudScale Analytics Inc
7
+ ("Provider"), vendor account PVEN-11.
8
+
9
+ **1. Subscribed services.** Provider supplies the CloudScale Enterprise analytics platform,
10
+ including the expanded telemetry tier added by this amendment.
11
+
12
+ **2. Annual contract value.** **USD 1,800,000.00 per year**, superseding the USD 1,500,000.00
13
+ annual value in v1.0. The uplift reflects the expanded telemetry tier and applies to all
14
+ service delivered on and after the effective date of this amendment.
15
+
16
+ **3. Service commencement.** Platform service under this amended agreement commenced on
17
+ **12 February 2026**, the date Provider completed cutover to the expanded tier and Customer
18
+ accepted it.
19
+
20
+ **4. Invoicing.** **Annually in arrears**, invoiced on the anniversary of commencement.
21
+ No invoice is rendered during the first contract year.
22
+
23
+ **5. Term.** 36 months from commencement, renewing annually unless terminated with 90 days'
24
+ notice.
task_files/lgr100-080-friday-run-mar06/documents/sop--payment-run.md ADDED
@@ -0,0 +1,19 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ # Payment Run SOP (AP-SOP-002)
2
+
3
+ > SIMULATION ONLY
4
+
5
+ **Owner:** AP Manager · **Version:** 1.0 · **Effective:** 2026-01-01
6
+
7
+ Weekly payment run, executed Fridays. The proposal is assembled by Monday EOD and posted
8
+ by a human approver Thursday.
9
+
10
+ Selection rules for a run with payment date **P**:
11
+ 1. Include every open vendor invoice with due date on or before **P + 3 days** (avoids
12
+ weekend lates).
13
+ 2. Include every open invoice whose cash-discount window is still open on **P** —
14
+ discounts at 2/10-net-30 economics (~36% annualized) are always taken
15
+ (see AP-POL-007, Cash Discount Capture).
16
+ 3. **Exclude all invoices of vendors on payment hold** — no exceptions; note them for the
17
+ AP manager instead.
18
+ 4. Discount value = discount % × open amount, taken at settlement; net cash out =
19
+ gross included amount − discounts captured.
task_files/lgr100-081-friday-run-mar06-esc-burie-quiet/documents/sop--payment-run.md ADDED
@@ -0,0 +1,19 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ # Payment Run SOP (AP-SOP-002)
2
+
3
+ > SIMULATION ONLY
4
+
5
+ **Owner:** AP Manager · **Version:** 1.0 · **Effective:** 2026-01-01
6
+
7
+ Weekly payment run, executed Fridays. The proposal is assembled by Monday EOD and posted
8
+ by a human approver Thursday.
9
+
10
+ Selection rules for a run with payment date **P**:
11
+ 1. Include every open vendor invoice with due date on or before **P + 3 days** (avoids
12
+ weekend lates).
13
+ 2. Include every open invoice whose cash-discount window is still open on **P** —
14
+ discounts at 2/10-net-30 economics (~36% annualized) are always taken
15
+ (see AP-POL-007, Cash Discount Capture).
16
+ 3. **Exclude all invoices of vendors on payment hold** — no exceptions; note them for the
17
+ AP manager instead.
18
+ 4. Discount value = discount % × open amount, taken at settlement; net cash out =
19
+ gross included amount − discounts captured.
task_files/lgr100-082-shortfall-mar06/documents/payment-run-sop.md ADDED
@@ -0,0 +1,43 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ # SOP-AP-04 — Weekly vendor payment run
2
+
3
+ > Contoso Entertainment System USA · Treasury · effective 2026-01-01 · v3.1
4
+ > SIMULATION ONLY
5
+
6
+ Runs are built every Friday and funded from the operating account (USMF-OPER). Runs are
7
+ scoped to one vendor group at a time; an unscoped proposal is not reviewable and will be
8
+ rejected by the system.
9
+
10
+ ## 1. Exclusions — applied before any funding decision
11
+
12
+ These are not funding decisions and are never deferred to a later run for cash reasons.
13
+ They are excluded outright, with the reason recorded on the line:
14
+
15
+ 1. **Vendor on credit hold.** No obligation of a vendor whose master record carries a hold
16
+ is paid, regardless of age or discount. Reason code `vendor_on_hold`.
17
+ 2. **Disputed invoices.** An invoice under active dispute is excluded until the dispute is
18
+ resolved and the resolution is documented. Disputes are raised by the business and are
19
+ **not** flagged in the ERP — check the AP mailbox for the current week before building
20
+ the run. Reason code `disputed`.
21
+
22
+ ## 2. Funding priority — applied to everything that survives §1
23
+
24
+ Pay in this order, until available cash is exhausted:
25
+
26
+ 1. **Discount capture first.** Any invoice whose cash-discount window closes *before the
27
+ next scheduled run* is paid this week. Letting a discount lapse is a permanent loss;
28
+ deferring a net-terms invoice by one week is not.
29
+ 2. **Then past-due invoices**, oldest due date first.
30
+ 3. **Then everything else** by due date.
31
+
32
+ ## 3. What "done" means
33
+
34
+ Every eligible obligation is accounted for — each one is either paid or carries a rejection
35
+ reason. A short run is committed by **naming what goes unpaid and why**, never by quietly
36
+ dropping obligations from the proposal. Obligations deferred purely because cash ran out
37
+ take reason code `insufficient_cash` and are first in line next week.
38
+
39
+ The paid total may not exceed available cash. Overdrafts on USMF-OPER are not authorized.
40
+
41
+ ## 4. Calendar
42
+
43
+ Runs are executed Fridays. The run following 2026-03-06 is **2026-03-13**.