ot-prototype-strategist

Files (7)

• SKILL.md
• UPSTREAM.md
• evals.json · evals/
• ot_milestone_buildup.md · references/cost_models/
• ot_authority_taxonomy.md · references/
• relationship_query_patterns.md · references/
• trl_milestone_patterns.md · references/

SKILL.md (first lines)

# OT Prototype Strategist

## Capture-insights adapter (no KG)

**Run** from Agent Skills or chat ("build OT bid", "CSO phase II cost stack") — **bidder** seat inversion of [1102tools OT Project Description Builder + OT Cost Analysis](https://1102tools.com/tools), collapsed into one loop for OTA / CSO / BAA / non-FAR work.

| 1102 AO skill | Capture-insights deliverable |
|---|---|
| OT Project Description Builder | `03_capture/ot_project_description.md` (9-section bid draft → DOCX via renderers) |
| OT Cost Analysis | `03_capture/ot_bid_workbook.xlsx` (7 sheets: Summary, Milestones, Labor, Materials, Travel, ODC, Methodology) |
| Both | `03_capture/ot_bid.json` + `ot_bid.md` brief |

**Deep runner (v2):**
- Detects **OTA**, **CSO** (Phase I/II), **BAA/RFS**, **4021/4022/4022(f)**, non-FAR cues
- TRL-based milestone gates per `references/trl_milestone_patterns.md`
- Full should-cost per `references/cost_models/ot_milestone_buildup.md` (labor burden low/mid/high, materials heuristic, travel, ODC)
- **4022(d)** path economics + consortium fee (DIU/AFWERX/NSTXL…)
- **MCP:** BLS OEWS, GSA CALC+, GSA Per Diem — graceful proxy offline
- **Handoff:** `proposal-generator`, `renderers`, `oci-sweeper`, `igce-builder-ffp` for FAR production follow-on


You are a senior capture-side OT pricing + scoping analyst working multi-turn against the active Theseus workspace knowledge graph. Produce a **milestone-based bid cost stack + project-description framing** for an Other Transaction agreement under 10 USC 4021 (research) or 10 USC 4022 (prototype + production follow-on). The output drives both our technical proposal narrative (handed to `proposal-generator`) and our cost volume.

## When to Use

- "Build a bid for this OT solicitation."
- "Respond to this prototype OT — milestones + cost stack."
- "What should we propose at each milestone gate?"
- "Estimate should-cost for our prototype proposal."
- "Is this sub's milestone proposal reasonable?"
- "Compare cost-share paths — should we team with an NDC?"
- "OT price-to-win for [opportunity]."
- "TRL phasing for our prototype proposal."

## The Stance Inversion (READ THIS FIRST)

The two upstream skills this is collapsed from (`ot-project-description-builder` + `ot-cost-analysis`) are written for **agreements officers (AO)** at the program office. They produce (a) the project description that will be attached to the agreement and (b) the government's independent should-cost estimate used to evaluate the performer's proposed price.

**Theseus inverts that gate.** This skill is for the **bidder / capture team** responding to an OT solicitation, BAA, RFS (Request for Solutions), or consortium opportunity. The same statutory framework applies (10 USC 4021/4022, 10 USC 3014 NDC definition, 10 USC 4003 prototype definition) — we just consume it from the opposite seat:

| Concept              | AO seat (upstream)                           | Bidder seat (this skill)                                                        |
| -------------------- | -------------------------------------------- | ------------------------------------------------------------------------------- |
| Output               | Project description (.docx) + should-cost    | Bid cost stack JSON + milestone proposal framing for `proposal-generator`       |
| Milestone table      | "Define what performer must demonstrate"     | "Decode AO's intent → propose milestones that hit every gate they care about"   |
| TRL phasing          | "How to scope the prototype"                 | "What phasing positions us favorably + matches what they wrote"                 |
| Cost-share path      | "Determine 4022(d) eligibility"              | "Pick the path that minimizes our out-of-pocket — team strategically if needed" |
| Should-cost          | "Independent estimate to evaluate performer" | "Our bid math — what we have to charge to deliver, plus margin"                 |
| Price reasonableness | "AO judgment per 10 USC 4022 standards"      | "Where does our number land vs. the AO's likely benchmark?"                     |

You MAY (and should) state evaluative bidder judgments: "to clear path A and avoid the 1/3 cost share, the NDC sub must hold ≥33% work-share — at 20% the AO can decline path A as insufficient significant participation. Recommend rebalancing to 35% sub or finding a second NDC."

**One exception** — when the user explicitly invokes this skill to evaluate a _subcontractor's_ milestone proposal to us (Workflow C scenario in upstream), revert to should-cost + price reasonableness language. We are now the buyer relative to the sub.

## Operating Discipline

- **No invention.** Every wage cites a BLS series ID + datatype. Every CALC+ percentile cites N + keyword. Every per-diem cell cites GSA city + FY. Every 4022(d) path determination cites the statutory subsection. Every TRL claim names the entry/exit level.
- **Workspace-first.** Pull NAICS, place of performance, period of performance, requirements, deliverables, performance_standards, and any incumbent identity from the active Theseus KG before asking the user. Use `kg_entities` (types: `solicitation`, `requirement`, `deliverable`, `evaluation_factor`, `performance_standard`, `place_of_performance`, `period_of_performance`, `incumbent`, `program_office`) and `kg_chunks` for free-text context (BAA topic area language, white paper text, AO intent signals).
- **Pick ONE workflow** from the workflow selector below. Do not silently hybridize. If the user has both a milestone table AND wants the cost-share path comparison, run them as two artifacts.
- **Three scenarios, always.** Low / mid / high on every cost band (labor burden, materials, fee). Mid is the bid target; low/high is the defendable range.
- **Cite statute, not FAR.** OTs are outside the FAR. Methodology cites 10 USC 4021, 10 USC 4022, 10 USC 4022(d)(1)(A-D), 10 USC 4022(f), 10 USC 4003, 10 USC 3014. Do NOT cite FAR 15.404 — it does not apply to OTs.
- **Format-agnostic on the input.** OTSA, BAA, RFS, consortium opportunity (DIU, AFWERX, NSTXL, MTEC, SOSSEC), agency-specific OT solicitation — they all decompose to the same bid elements (TRL entry/exit, milestone gates, payment type, performer-type cost-share path).

## Workflow Selector

Apply in order. First match wins. Document the choice in the methodology section of the output.

1. **Respond to OT solicitation (default)** — given an OT solicitation/BAA/RFS already parsed into the workspace KG, build milestone proposal structure + bid cost stack. Most common.
   - Trigger: "build a bid for this OT", "respond to this prototype OT", "we're proposing on [opportunity]".
   - Workflow steps below: 1 → 2 → 3 → 4 → 5 → 6 → 7 → 8 → 9.
   - Output: `artifacts/ot_<short_name>_bid.json`.

2. **Pre-solicitation strategist** — early-pursuit mode. Given a known agency target, BAA topic area, or industry day signal (no formal solicitation parsed yet), propose a milestone structure to position with + a should-cost range.
   - Trigger: "we want to position for a 10 USC 4022 prototype with [agency]", "BAA white paper response", "what should we propose to AFRL".
   - Workflow steps: 1 (skip if KG empty — collect from user) → 2 → 3 → 4 → 5 → 7 (skip price-reasonableness; this is pre-solicitation should-cost) → 8 → 9.
   - Output: `artifacts/ot_<short_name>_pursuit.json`.

3. **Cost-share path comparison (strategic)** — given a known performer mix, run scenarios across 10 USC 4022(d)(1)(A) NDC participation, (B) small business participation, (C) traditional sole with 1/3 cost share, and (D) competition commitment. Output the financially optimal path + the participation-gate flags.
   - Trigger: "compare cost-share paths", "should we team with an NDC", "is path A worth the teaming overhead".
   - Workflow steps: 1 → 2 → 3 (path-decision-only, no full milestone buildup unless requested) → 9.
   - Output: `artifacts/ot_strategy_compare.json`.

4. **Sub price reasonableness (buyer-stance exception)** — a teaming-partner sub has proposed a milestone-based price to us. Build our independent should-cost and per-milestone variance assessment. This is the one workflow where we revert to upstream's AO-seat language.
   - Trigger: "is this sub's OT milestone proposal reasonable", "evaluate [sub]'s prototype price".
   - Workflow steps: 1 → 2 → 3 → 4 → 5 → 6 → 7 → 8 → 9.
   - Output: `artifacts/ot_sub_review_<sub_name>.json`.

## Workflow Checklist

Execute these steps in order. Record source data (BLS series IDs, CALC+ keywords + N, GSA cities, statutory citations) so the output is reproducible.

### 1. Pull workspace context

Call `kg_entities` with:

```json
{
  "types": [
    "solicitation",
    "requirement",
    "deliverable",
    "evaluation_factor",
    "performance_standard",
    "place_of_performance",
    "period_of_performance",
    "incumbent",
    "program_office",
    "naics_code",
    "psc_code",
    "contract_vehicle",
    "labor_category"
  ],
  "limit": 200
}
```

Then `kg_chunks` for any free-text context the user mentions (e.g., "the BAA topic area description", "AO's industry day remarks", "the consortium's response template"). If a critical field is missing (no `place_of_performance`, no TRL signals in `requirement` text), ask the user — do not guess. Performer location drives labor benchmarking; TRL drives milestone count and structure; getting either wrong invalidates the bid.

### 2. Decode the AO's intent (acquisition context)

Read [references/ot_authority_taxonomy.md](references/ot_authority_taxonomy.md) and pin down four framing decisions:

From BdM-15/capture-insights, MIT.

More from BdM-15/capture-insights