qwen3.6-35b-a3b on “A digest of several sources”: with Minto, worse on quality, structure unchanged

Structure 6 → 6 of 8 0 Quality 9 → 8 of 10 −1

The judge on the output with Minto

The opening action is strong, while the older package commitment is blurred and the diagram date is attached to the wrong sources.

Mode digest · case 09-sources-digest · run v1.7.0-final

Structure of 8, 2 points per axis
Structure	Without Minto 6	With Minto 6
Top	2	2
Key-line composition	1	1
Levels	1	1
Order and kind	2	2

Quality of 10, 2 points per axis
Quality	Without Minto 9	With Minto 8	Change
Top	1	2	+1
Same-kind grouping	2	1	−1
Explainable order	2	2
MECE	2	1	−1
Visible, proportionate structure	2	2

Outputs and the judge’s words are quoted as written, in English.

Without Minto With Minto

Without Minto

control prompt · blind label out-07

invented_facts
quality after the rubric’s penalty 7/10

It is concise and actionable, but buries Fernway’s role in Q7 and promises more closure than the evidence supports.

Bottom Line Up Front
We can resolve the compliance blockers today by sending the draft architecture diagram to unlock the remaining data-security answers, while formally escalating the overdue package list and remote-work clarifications. This will allow us to fix the build-phase pricing.

What Is Already Confirmed
• Pre-production review: Mandatory. A penetration test may also be required, but its scope will be determined during the review itself.
• Hardening standards: An internal framework exists; conformance does not waive the review requirement.
• Open-source dependencies: A policy is in place. Usage is assessed case-by-case during the security review, with no standard turnaround time.
• Remote work: Staffing from Ireland is explicitly approved.

What Still Requires Aldergate Action
• Approved package list: Overdue since 12 May (last update 9 July). Needed to validate open-source dependencies against the policy.
• Data storage, encryption & logging specs: Requirements exist but cannot be confirmed until reviewed against an architecture diagram.
• Lisbon-based engineers: Two Fernway staff are based in Portugal; Aldergate has not yet responded to this inquiry.
• Direct compliance channel: All current communication is routed through delivery management. We need direct access to S. Okafor (Information Security Owner) to accelerate approvals.

Recommended Actions for Today

Send the draft architecture diagram to Aldergate immediately to trigger the storage/encryption/log answers.
Escalate the approved-package list and request explicit confirmation on the Lisbon engineers.
Request a direct introduction to S. Okafor to bypass the delivery-only handoff and secure timely dependency approvals.

Once these items are addressed, the security requirements will be fully confirmed, allowing us to lock the build-phase budget.

With Minto

prompt with the skill · blind label out-08

invented_facts
quality after the rubric’s penalty 7/10

The opening action is strong, while the older package commitment is blurred and the diagram date is attached to the wrong sources.

Submit our draft architecture diagram and formally request the overdue approved package list and remote-work confirmation from Aldergate’s security team to unblock build-phase pricing today.

We have been waiting on Aldergate’s compliance office to finalize security requirements so we can fix the build-phase price. They have confirmed that a security review is mandatory and that an internal hardening framework exists, but several critical dependencies remain stalled: the approved package list is over two months late, data storage and encryption details cannot be provided until we share our architecture, and the status of our Lisbon-based engineers is still unknown. To move forward, we need to address these specific gaps immediately.

Requirements already confirmed

A security review is mandatory before production, independent of any hardening framework conformance [Source 1].
Critical dependencies will be assessed case-by-case during the review rather than pre-approved [Source 1].
Remote work from Ireland is permitted; requests for other countries remain unapproved [Source 1].

Outstanding items requiring immediate action

Send the draft architecture diagram (completed 11 Aug) to Aldergate so they can specify data storage, encryption, and logging requirements [Source 1, Source 2].
Contact S. Okafor (Aldergate information security owner) directly to obtain the approved package list and clarify the remote-work policy for Lisbon-based staff [Source 2, Source 3].
Close the open compliance risks and lock the build-phase price once the diagram is submitted and the package list is received [Source 3].

The judge

Recurring defect	without	with Minto
Unsupported facts, causal links, or closure conditions are introducedthe skill has the rule; the model did not follow it · Core loop §8.4: every element rests on source material	present	present
Fernway’s Q7 move is not cleanly separated from what Aldergate owesthe rule is there but buried · Mode 4: cross-source facts must be combined and grouped by the reader’s question	present	absent
Lisbon is converted to approval/geography or omittedthe skill has the rule; the model did not follow it · Mode 4: action-changing facts must survive compression	present	absent
Load-bearing claims lack complete source locatorsthe skill has no such rule · Mode 4: source reference on every load-bearing claim	present	absent
The top overstates ownership or the conditions for fixing pricingthe skill has the rule; the model did not follow it · Core loop §2 and Mode 4 opening answer	present	absent

Decisive questions

Does the first line answer what the tech lead should do, or does it describe the sources?
without Yes, but it overpromises that today’s actions will resolve pricing.
with Minto Yes; the first sentence gives the immediate actions.
Is the digest organized by answer (settled / blocked on us / owed by them or equivalent) or source by source?
without Answer-organized, though Q7 is placed under Aldergate action rather than Fernway’s blocker.
with Minto Answer-organized.
Is the package-list commitment dated from 12 May (98 days) or from 9 July?
without Dates it from 12 May but omits the 98-day age.
with Minto Neither exact alternative: it says only “over two months late.”
Are the Q7 refusal and the unsent 11 August draft combined into one point about Fernway's own move?
without Yes.
with Minto Yes.
How is Lisbon handled: unresolved exposure, presented as approved, or silently dropped?
without Correctly unresolved, but converted to “Portugal.”
with Minto Correctly unresolved.
Do load-bearing claims carry a source reference?
without No.
with Minto Yes, but the 11 August diagram claim is cited to the wrong sources.
Which of these appear: marker legend, mermaid, findings table, meta-comment about group order?
without None.
with Minto None.
Is the length within roughly one chat screen?
without Yes.
with Minto Yes.
Which facts, dates, or commitments were invented?
without It adds Portugal, guarantees pricing closure, and claims direct Okafor access will secure timely approvals.
with Minto It claims the diagram and package list will close the risks and lock pricing, although other security decisions remain open.

Judge’s check	without	with Minto
Answer first	yes	yes
First-level elements	3	2
First-level kind matches the gold	no	yes
Invented facts	yes	yes
Mode respected	yes	yes
SCQ introduction present	yes	yes

3 more judge checks do not apply to mode digest

How the test was built

qwen3.6-35b-a3b received the same task twice. The control prompt is one framing line and the task. The prompt with Minto is the same task preceded by the full text of the skill. The judge read both outputs among 16 under blind labels and did not know which one had the skill.

Prompt without Minto 3 667 B

You are working on a business document using the Minto Pyramid Principle.

<!-- Original fixture. The scenario, company, people, systems and numbers are
     invented for this benchmark. The defect pattern it tests — a digest of several
     read sources delivered answer-first into chat — is modeled on real usage. -->

# Fixture 09: sources digest

**Mode:** `digest`
**Language:** `en`

## Context

You are assisting the tech lead of Fernway Systems, a vendor building a customer
support assistant for Aldergate Bank. Before Fernway can price the build phase, the
bank's compliance office must confirm its security requirements. You have just read
the three sources below in full. The tech lead has not read them and will not; they
are deciding what to do about the compliance track today.

## Before

Source 1 — `gate-review.xlsx`, sheet "Security questions", filled in by Aldergate's
compliance office on 12 May:

```text
Q1 Is a security review required before production?
   A: Yes. A review is mandatory; whether a penetration test is also needed is
   decided during the review itself, by the assigned reviewer.
Q2 Is there a security standard or checklist we can build against?
   A: An internal hardening framework exists; a review is still required
   regardless of conformance.
Q3 What is the policy on open-source dependencies?
   A: A policy exists. The approved-package list will be sent to Fernway.
   Owner: D. Mercer. Date: 12 May.
Q4 Must critical dependencies be approved individually?
   A: Dependency usage is assessed during the security review.
Q5 What is the turnaround time for dependency approval?
   A: There is no standard turnaround; it is case by case.
Q6 Can Fernway staff work from outside the UK?
   A: Work from Ireland is acceptable. (No other country is mentioned.)
Q7 Where may customer data be stored, and with what encryption and logging?
   A: Declined — "cannot be answered without an architecture diagram and a
   design walkthrough". Encryption and logging requirements exist but are
   not applicable outside an architectural context.
```

Source 2 — follow-up note from the 9 July delivery sync, `sync-0709.md`:

```text
Attendees: R. Whitcombe (Aldergate, delivery), tech lead (Fernway), PM (Fernway).
- Whitcombe repeated that the approved-package list "is being prepared" and will
  reach Fernway "shortly". Action on Whitcombe, no date attached.
- Compliance contact named for the first time: S. Okafor, information security
  owner. Okafor has not attended any joint meeting so far; Whitcombe passes
  questions along.
- Agreed: Fernway sends the draft architecture diagram before the next sync so
  the storage and encryption questions can get a real answer.
- Fernway's PM noted two Fernway engineers are based in Lisbon. Nobody from
  Aldergate responded to the remark in the meeting.
```

Source 3 — Fernway's internal risk register, rows touching compliance, as of
18 August:

```text
R-07  Security requirements unconfirmed; build-phase price cannot be fixed.
      Owner: tech lead. Status: open since 3 May.
R-11  Approved-package list not received. First promised 12 May (Mercer, per
      gate review), promised again 9 July (Whitcombe). Status: open, 98 days.
R-14  Architecture diagram not yet sent to Aldergate; blocks the data-storage
      answer (gate review Q7). Draft exists since 11 August. Status: open.
```

## Task

Give the tech lead a digest of these three sources in chat: what is already
answered on the compliance track, and what still has to be obtained from
Aldergate. The tech lead will act on it today.

Return only the result, with no explanation of how you produced it.

Prompt with Minto five parts, in order · 31 240 B

Below is a skill written as an instruction. Read it in full and apply it to the task at the end.
===== SKILL.md ===== SKILL.md at f4813ce 17 235 B
===== references/rules.md ===== references/rules.md at f4813ce 6 464 B
===== references/templates.md ===== references/templates.md at f4813ce 3 742 B
===== TASK ===== the whole case input
Return only the result, with no explanation of how you produced it.

Case input before.md

Fixture 09: sources digest

Mode: digest
Language: en

Context

You are assisting the tech lead of Fernway Systems, a vendor building a customer
support assistant for Aldergate Bank. Before Fernway can price the build phase, the
bank's compliance office must confirm its security requirements. You have just read
the three sources below in full. The tech lead has not read them and will not; they
are deciding what to do about the compliance track today.

Before

Source 1 — gate-review.xlsx, sheet "Security questions", filled in by Aldergate's
compliance office on 12 May:

Q1 Is a security review required before production?
                   A: Yes. A review is mandatory; whether a penetration test is also needed is
                   decided during the review itself, by the assigned reviewer.
                Q2 Is there a security standard or checklist we can build against?
                   A: An internal hardening framework exists; a review is still required
                   regardless of conformance.
                Q3 What is the policy on open-source dependencies?
                   A: A policy exists. The approved-package list will be sent to Fernway.
                   Owner: D. Mercer. Date: 12 May.
                Q4 Must critical dependencies be approved individually?
                   A: Dependency usage is assessed during the security review.
                Q5 What is the turnaround time for dependency approval?
                   A: There is no standard turnaround; it is case by case.
                Q6 Can Fernway staff work from outside the UK?
                   A: Work from Ireland is acceptable. (No other country is mentioned.)
                Q7 Where may customer data be stored, and with what encryption and logging?
                   A: Declined — "cannot be answered without an architecture diagram and a
                   design walkthrough". Encryption and logging requirements exist but are
                   not applicable outside an architectural context.

Source 2 — follow-up note from the 9 July delivery sync, sync-0709.md:

Attendees: R. Whitcombe (Aldergate, delivery), tech lead (Fernway), PM (Fernway).
                - Whitcombe repeated that the approved-package list "is being prepared" and will
                  reach Fernway "shortly". Action on Whitcombe, no date attached.
                - Compliance contact named for the first time: S. Okafor, information security
                  owner. Okafor has not attended any joint meeting so far; Whitcombe passes
                  questions along.
                - Agreed: Fernway sends the draft architecture diagram before the next sync so
                  the storage and encryption questions can get a real answer.
                - Fernway's PM noted two Fernway engineers are based in Lisbon. Nobody from
                  Aldergate responded to the remark in the meeting.

Source 3 — Fernway's internal risk register, rows touching compliance, as of
18 August:

R-07  Security requirements unconfirmed; build-phase price cannot be fixed.
                      Owner: tech lead. Status: open since 3 May.
                R-11  Approved-package list not received. First promised 12 May (Mercer, per
                      gate review), promised again 9 July (Whitcombe). Status: open, 98 days.
                R-14  Architecture diagram not yet sent to Aldergate; blocks the data-storage
                      answer (gate review Q7). Draft exists since 11 August. Status: open.

Task

Give the tech lead a digest of these three sources in chat: what is already
answered on the compliance track, and what still has to be obtained from
Aldergate. The tech lead will act on it today.

The judge’s gold gold.md

Gold 09: sources digest

Original gold for an invented scenario.

Expected structure

Top: an answer to the tech lead's question, not a description of the sources.
The strongest available top is an action with its ground: the blocking questions
cannot be answered by more reading — send the architecture diagram (drafted and
sitting unsent since 11 August) and escalate the approved-package list as an
obligation now 98 days old. A top that merely states "five of seven questions are
answered" is weaker but acceptable if the two actions lead the groups.

Three same-kind groups the material supports, in any defensible order:

Settled — do not re-ask. Review is mandatory (pen test decided during it);
a hardening framework exists but exempts nothing; dependencies are assessed in
the review with no standard turnaround; Ireland is cleared for remote work.
Blocked on Fernway's own move. The storage/encryption/logging answer (Q7)
was declined pending an architecture diagram; the draft has existed since
11 August and has not been sent (R-14). This is Fernway's blocker, not
Aldergate's.
Owed by Aldergate and overdue. The approved-package list, first promised
12 May by Mercer per the gate review — not 9 July by Whitcombe, so 98 days
overdue, which makes the ask an escalation of an old commitment rather than a
reminder. Direct access to Okafor, who has never attended and is reachable
only through Whitcombe.

Cross-source synthesis the digest must make

The package-list promise appears in two sources with two dates; the digest
must date it from 12 May (the older commitment), not 9 July.
Q7's "declined" plus R-14's "draft exists since 11 August" must combine into
"the ball is in Fernway's court"; either fact alone misses the point.
The Lisbon remark: Q6 clears Ireland only. Two engineers in Lisbon are an
unanswered exposure, not a settled fact. Presenting Lisbon as approved is an
invention; omitting it entirely loses the one new risk the sources surface.

Presentation

Chat prose for one reader: answer first, grouped support, a reference to which
source each load-bearing claim comes from (sheet, note, or register row —
exact citation format is free). No marker legend, no mermaid diagram, no
findings table, no meta-commentary about how the groups are ordered. Headings or
bold group labels are acceptable; a full document apparatus (SCQ labels as
headings, scoring, appendix) is excessive. Length: roughly one chat screen —
condensing the six settled answers into one or two lines each is proportional
omission, not source loss.

Judge gpt-5.6-sol, effort high · skill from commit f4813ce · run v1.7.0-final