Value Engineering Advisor.

Overview

Review a value-engineering (VE) list submitted by a CM, GC, or owner — typically post-bid, late-DD, or any moment when the project budget has slipped — and produce a categorized recommendation memo per item. Output lives in docs/value-engineering/. Each item is classified ✅ ACCEPT / ⚠️ ALTERNATE / ❌ BLOCK with a one-line "why" naming the design-intent, performance, lifecycle, code, schedule, or risk-transfer concern.

This skill is for the architect / AOR / owner's-rep perspective. The memo is a recommendation that the AOR, owner, and CM/GC discuss at the next VE workshop or OAC meeting; it is not a binding directive and does not replace the AOR's stamp on subsequent CDs / ASIs / change orders.

Aligned with ASTM E1699 (VE Job Plan), CSI MasterFormat 2020 (cuts organized by division), AIA A201-2017 §7 (Changes in the Work) + G701 (Change Order), AIA B101-2017 §3.6.1.5 (architect's cost review), CMAA preconstruction services scope.

When to use

Trigger on:

• Pasted VE lists (CSI division, item, current spec, proposed substitution, $ savings).
• Project file paths to a VE workshop package.
• Explicit asks: "VE review", "value engineering", "where can we cut", "review VE round 2", "what should we accept from the CM's VE list".
• Hand-off from bid-review when the low bid exceeds budget or qualifications include "VE assumed".

Do NOT trigger for:

• Cost estimating from scratch (this skill assesses proposed cuts; it does not generate $ values — use the CM/GC quote).
• Constructability review (use constructability-review).
• Bid leveling (use bid-review).
• Spec-section completeness review (use spec-review).
• Final code / egress / energy check after VE accepted (use code-review).

Workflow

Step 1 — Bootstrap if needed

Check that docs/value-engineering/index.md exists. If not:

python scripts/init_ve_workspace.py

Step 2 — Anchor the project context

Read docs/project.md if it exists. Pull:

• Phase (SD / DD / CD / post-bid / CA — the later the phase, the more expensive the redesign triggered by an accepted cut).
• Delivery method (DBB / CMAR / DB / IPD — affects who carries VE risk and who proposes alternates).
• Budget gap ($ over budget; the size of the gap drives how aggressive cuts must be).
• Owner type (institutional / commercial / public / healthcare / K-12 — affects risk tolerance and approval process; e.g., CA HCAI rejects many envelope cuts, NYC SCA rejects many finish cuts).
• Project type (the heuristics in references/csi_division_heuristics.md distinguish public/identity vs. back-of-house — a lobby cut hits differently than a mech-room cut).
• Code / certification anchors (LEED Gold pursued? Passive House? LL97 carbon cap? IECC stretch? Each constrains what can be cut without losing the target).
• Named AHJ (NYC DOB, FDNY, CA HCAI, DSA, GSA, USACE, FEMA region, state DEP/DEC — the AHJ controls what alternates require re-review).

Step 3 — Apply per-item assessment

For each VE item, work through the six-axis decision tree in references/ve_assessment_axes.md:

1. Design intent impact (concept identity, public vs. back-of-house).
2. Performance / code impact (envelope R-value, fire rating, STC, ADA, IECC, NFPA).
3. Lifecycle cost (upfront savings vs. O&M penalty over 20–30 yr).
4. Schedule impact (does the cut actually save schedule, or trigger redesign + re-permit + AHJ re-review?).
5. Risk transfer (does the cut shift risk back to owner — e.g., reducing waterproofing transfers leak risk).
6. Reversibility (can this be reinstated later in CA, or once gone is it gone?).

For division-specific cuts, consult references/csi_division_heuristics.md keyed to the CSI MasterFormat division. Heuristics name the typical impact — but the project-context anchors from Step 2 override generic heuristics.

Classify per the rule:

• ✅ ACCEPT if design-neutral + no performance hit + lifecycle neutral-or-positive + no risk transfer.
• ⚠️ ALTERNATE if minor impact on one axis — propose a modified approach (e.g., "specify equivalent product line by alt manufacturer" or "reduce spec rather than eliminate").
• ❌ BLOCK if any of: kills design intent in a public/identity space, code or performance failure, unacceptable risk transfer, irreversible loss, or voids a tracked certification target without owner sign-off.

Each item gets a one-line "why" that names the axis it failed/passed and cites the reference (e.g., csi_division_heuristics.md §07 27 26 or IBC §720 for fire rating).

Step 4 — Aggregate into the recommendation memo

Use the template at assets/ve_memo_template.md. The memo includes:

• Header — project, phase, VE round, date, gap $ vs. accepted savings $.
• Categorized table — three sections (Accept / Alternate / Block) with per-item one-liners and citation column.
• Downstream impacts — follow-on actions: spec revisions (Section #), drawing revisions (sheet #), schedule impact (days), code re-review needed (egress / energy / accessibility / fire), CA-phase tracking (LEED point loss, warranty length change, certification target risk).
• Open questions — items that need owner / AHJ input before classifying (e.g., "is LEED Gold still hard target if MR credit drops 2 pts?").

Write the memo to docs/value-engineering/<round-number>-<YYYY-MM-DD>.md and append a row to docs/value-engineering/index.md.

Step 5 — Hand off

The memo is a recommendation, not a directive. Hand off to:

• AOR for stamp on accepted items. Any ⚠️ accepted as alternate must roll into ASI per AIA G710 or CCD per AIA G714.
• spec-review to verify spec-section consistency on accepted alternates (especially Part 2 product/manufacturer changes).
• code-review if any accepted cut touches egress, fire-resistance, accessibility, or envelope thermal.
• constructability-review if any accepted alternate changes sequencing, trade interface, or installation tolerance.
• submittal-review later in CA when the substituted product's submittal arrives.

Anti-patterns

• Never invent $ savings. Use only what the CM/GC quoted. If a number is missing, mark "TBD — request from CM" — do not estimate.
• Never accept a cut that drops below code minimum. Even if "the AHJ has approved similar," the AOR carries the stamp; default to ❌ and require AHJ alt-means letter on file before re-classifying.
• Never recommend a cut that voids a tracked certification target (LEED, WELL, PHIUS, Title 24 above-code, LL97 cap) without explicit owner sign-off on the loss. The memo names the certification impact in the "downstream" column.
• Never present this skill's six-axis framework as a published industry standard. ASTM E1699 publishes the VE Job Plan (Information → Function → Creative → Evaluation → Development → Presentation); the six-axis classification used here is a structured review heuristic synthesized to surface obligations from ASTM E1699, AIA contracts, and CSI MasterFormat.
• Never skip the "Reversibility" axis on irreversible scope (envelope assemblies, structural waterproofing, in-slab radiant, embedded blocking). Once gone, retrofit cost typically exceeds 5–10× the original install cost.

Resources

• scripts/init_ve_workspace.py — bootstrap docs/value-engineering/ and index.md.
• references/ve_assessment_axes.md — six-axis decision framework with sub-criteria, common failure modes, example findings.
• references/csi_division_heuristics.md — CSI MasterFormat 2020 division-keyed cuts with design / performance / lifecycle / risk impact heuristics.
• references/astm_e1699_methodology.md — VE Job Plan outline (six phases) for context on industry-standard methodology.
• assets/ve_memo_template.md — recommendation memo template for docs/value-engineering/.