BdM-15/capture-insights / skills/ot-prototype-strategist
ot-prototype-strategist
VerifiedFederal Other Transaction (OT) prototype bid strategist for 10 USC 4021 research and 10 USC 4022 prototype/4022(f) production-follow-on agreements. USE WHEN the user asks to "build an OT bid", "respond to this OT solicitation", "estimate should-cost for an OT prototype", "is this milestone price reasonable", "compare 4022(d) cost-share paths", "team with an NDC to avoid 1/3 cost share", "OT price-to-win", or "TRL milestone phasing" — any milestone-based prototype scoping, OT cost stack, or 4022(d) cost-share strategy question. Reconstructs the AO's TRL phasing from the active workspace KG, picks the 4022(d) path that minimizes our exposure, builds a per-milestone cost stack from BLS OEWS + GSA CALC+ + GSA Per Diem, and emits a JSON envelope handed to `proposal-generator`. DO NOT USE FOR FAR-based contracts (`price-to-win`/`proposal-generator`), incumbent research (`competitive-intel`), FAR clause audit (`compliance-auditor`), or prime-to-sub SOW drafting (`subcontractor-sow-builder`).
| File | Kind | Size |
|---|---|---|
| SKILL.md | skill md | 23.1 KB |
| UPSTREAM.md | markdown | 7.2 KB |
| evals/evals.json | config | 9.6 KB |
| references/cost_models/ot_milestone_buildup.md | markdown | 10.2 KB |
| references/ot_authority_taxonomy.md | markdown | 7.7 KB |
| references/relationship_query_patterns.md | markdown | 4.8 KB |
| references/trl_milestone_patterns.md | markdown | 7.4 KB |
evals/evals.json
View on GitHub{
"skill_name": "ot-prototype-strategist",
"evals": [
{
"id": 1,
"prompt": "We're an NDC startup pursuing a TRL 4-6 cyber prototype OT under 10 USC 4022. The solicitation is in the workspace (DIU consortium-brokered, San Diego, ~18 months, target $2.4M). Build the milestone proposal structure AND our bid cost stack so I can see what to propose at each gate and what it should cost us to deliver.",
"expected_output": "A JSON envelope at artifacts/ot_prototype_bid.json with workflow='respond_to_ot_solicitation', authority='10 USC 4022', performer_type='ndc', cost_share_path='4022(d)(1)(A)' with required=false, milestones[] (3-5 entries with milestone_id, phase, trl_in/trl_out, est_duration_months, deliverables[], payment_type, exit_criterion, should_cost_mid), labor[] benchmarked from BLS for the SOC codes (cybersecurity engineer 15-1212 minimum) at the San Diego MSA with explicit aging factor, calc_plus[] anchors with N reported, materials[] sized to prototype-type heuristic (software/hybrid 10-40%), travel[] for any test-event trips with GSA per diem, consortium_fee block (5% default for DIU), totals.{government_obligation_low/mid/high, performer_share}, and a bid_recommendation block with target_price, rationale, and at least 3 risk_flags. Methodology cites 10 USC 4022, 10 USC 4022(d)(1)(A), 10 USC 4003.",
"expectations": [
"The skill called kg_entities at least once requesting types including 'solicitation', 'requirement', 'deliverable', 'evaluation_factor', 'performance_standard', or 'place_of_performance'.",
"The skill called mcp__bls_oews__get_wage_data or mcp__bls_oews__igce_wage_benchmark for cybersecurity SOC 15-1212 (or 15-1252 software developer) at the San Diego metro.",
"The skill called mcp__gsa_calc__igce_benchmark or mcp__gsa_calc__price_reasonableness_check at least once.",
"The skill called mcp__gsa_perdiem__get_per_diem_by_zip or mcp__gsa_perdiem__estimate_travel_cost for the test-event travel line item (San Diego or test range).",
"The skill read references/ot_authority_taxonomy.md via read_file.",
"The skill read references/trl_milestone_patterns.md via read_file.",
"The skill read references/cost_models/ot_milestone_buildup.md via read_file.",
"The skill read references/relationship_query_patterns.md via read_file before calling kg_entities.",
"The skill called write_file with target path 'artifacts/ot_prototype_bid.json'.",
"The output authority field equals '10 USC 4022' (NOT '10 USC 4021' — this is a prototype, not research).",
"performer_type equals 'ndc' AND cost_share_path equals '4022(d)(1)(A)' AND cost_share_required equals false.",
"milestones array has at least 3 entries; every entry has a non-null exit_criterion (binary pass/fail); every payment_type is 'fixed' or 'cost-type' (not 'tbd').",
"Every labor entry has a buildup with non-null hourly rate AND the BLS series_id is reported.",
"consortium_fee.rate equals 0.05 (DIU default) and is added to government_obligation, not deducted from performer share.",
"bid_recommendation has a non-null target_price AND at least 3 risk_flags.",
"Methodology citations include both '10 USC 4022' AND '10 USC 4022(d)(1)(A)'."
]
},
{
"id": 2,
"prompt": "An AFRL research OT under 10 USC 4021 just dropped — TRL 2-4 autonomy algorithm work, university PI, Boston metro, 24 months, $1.8M. We're a traditional prime teaming with MIT. Build the project description framing AND the bid cost stack — but get the cost-share story right because the source doc shows confusing numbers.",
"expected_output": "A JSON envelope at artifacts/ot_research_bid.json with workflow='respond_to_ot_solicitation', authority='10 USC 4021', cost_share_required=false (4021 has NO statutory cost-share trigger), milestone structure mapped to research phases (typically 2-3 milestones across TRL 2-4), labor[] using academic burden scenarios (1.65/1.85/2.05 instead of 1.8/2.0/2.2) and grad_research_assistant entries at $55-75/hr institutional rate (NOT BLS-derived), postdoc entries at $80-110/hr, materials sized to research heuristic (low — 10-20%), and methodology that EXPLICITLY states '10 USC 4022(d) cost-sharing paths are statutorily inapplicable' to a 4021 research OT.",
"expectations": [
"The skill called kg_entities for solicitation/requirement/place_of_performance.",
"The skill called mcp__bls_oews__get_wage_data for the Boston metro for at least one labor category.",
"The skill read references/ot_authority_taxonomy.md AND references/cost_models/ot_milestone_buildup.md via read_file.",
"The skill called write_file with target path 'artifacts/ot_research_bid.json'.",
"The output authority field equals '10 USC 4021' (NOT '10 USC 4022').",
"cost_share_required equals false AND methodology_notes contains the phrase 'statutorily inapplicable' or '4021' adjacent to 'no cost-share'.",
"At least one labor entry uses an academic burden scenario (low/mid/high in the 1.65/1.85/2.05 band, not 1.8/2.0/2.2).",
"At least one labor entry has role='graduate_research_assistant' OR role='grad_ra' with hourly rate between 55 and 75 (institutional rate, NOT BLS), AND methodology_notes flags the institutional-rate basis (not BLS-derived).",
"Methodology citations include '10 USC 4021' but do NOT cite '10 USC 4022(d)' as a binding authority on this agreement."
]
},
{
"id": 3,
"prompt": "Strategic question — we're a traditional prime (no NDC, no SB) with no teaming partner yet, looking at a 10 USC 4022 prototype OT. We can either bid solo (path C, 1/3 cost share required) or team with a small NDC sub. Run the cost-share scenario comparison so we know which path is cheaper out-of-pocket. Assume a $3.0M total project, traditional prime carries 80% of the work, NDC sub carries 20%.",
"expected_output": "A JSON envelope at artifacts/ot_strategy_compare.json with workflow='cost_share_path_compare', two scenarios labeled scenario_solo (path 4022(d)(1)(C), 1/3 cost share) and scenario_team (path 4022(d)(1)(A), NDC participation satisfies, no cost share if NDC participation is 'significant' = ≥33%), per-scenario government_obligation_mid and performer_out_of_pocket_mid, a participation_gate flag noting that 20% sub work-share is BELOW the typical 33% significant-participation threshold and that path A is at AO discretion, and a recommendation that increasing sub work-share to ≥33% OR adding additional NDC participation reduces our exposure by ~$1M at the standard 1/3 share. Methodology cites 10 USC 4022(d)(1)(A), 10 USC 4022(d)(1)(C), and 10 USC 3014 (NDC definition).",
"expectations": [
"The skill called kg_entities for solicitation/requirement/place_of_performance/incumbent if present.",
"The skill read references/ot_authority_taxonomy.md via read_file.",
"The skill called write_file with target path 'artifacts/ot_strategy_compare.json'.",
"The output workflow field equals 'cost_share_path_compare'.",
"The output has TWO labeled scenarios: scenario_solo (or path_c) AND scenario_team (or path_a).",
"scenario_solo.cost_share_path equals '4022(d)(1)(C)' AND scenario_solo.performer_share_pct equals approximately 0.333 (within 0.01 tolerance).",
"scenario_team.cost_share_path equals '4022(d)(1)(A)' AND scenario_team.performer_share_pct equals 0.0.",
"The output contains a participation_gate flag noting that 20% sub work-share is below the 33% significant-participation threshold OR that 4022(d)(1)(A) significant-participation is an AO judgment without a statutory percentage.",
"Methodology citations include '10 USC 4022(d)(1)(A)' AND '10 USC 4022(d)(1)(C)' AND '10 USC 3014'."
]
},
{
"id": 4,
"prompt": "We're an NDC startup pursuing a TRL 4-6 cyber prototype OT under 10 USC 4022. The solicitation is in the workspace. Build the milestone proposal structure AND our bid cost stack, then return the standard narrative summary.",
"expected_output": "Narrative summary covering the 7 numbered points in the SKILL.md (workflow chosen, authority + performer + cost-share path, milestone structure, totals, top 3 risk flags, hand-off, statutory citations), backed by an actually-written JSON artifact on disk. Cover-note prose may use bold-bullet stats, milestone field readouts (Deliverables:, Exit criterion:, Traces to:, TRL in/out:), structured-field disclosure of artifact keys, and Class B reasoning markers (default, standard per, per AO convention, per taxonomy reference) — substantive claims live in the JSON envelope's source_chunks fields.",
"expectations": [
"The skill MUST call write_file with a target path matching artifacts/ot_*.json before drafting the narrative summary (per SKILL.md v0.2.0 'REQUIRED — DO NOT SKIP' directive).",
"tools/audit_skill_grounding.py --skill ot-prototype-strategist reports grounding_ratio is >= 0.85 (milestone-and-cost-stack reasoning skill; the JSON artifact's source_chunks fields are the citation-of-record, the narrative is the human-readable distillation; 147.10 calibration credits cost-stack labor/travel bullets, BLS vintage caveats, Government-obligation envelope readouts, and capability-pointer offers as exempt).",
"tools/audit_skill_grounding.py reports claims_exempt is at least 0.40 of claims_total (most cover-note prose should be recognized as structural / milestone-bullet / structured-field / reasoning-leap, not unsourced claims).",
"No regression below 0.30 across at least three consecutive runs of the same prompt."
]
}
]
}
Related skills
| Skill | Stars | Updated | Works with | Status | Install |
|---|---|---|---|---|---|
| rea-analysis-optimizationmorluto/reaOptimize REA providers and shared analysis workflows for agent context, latency, CPU, and memory while preserving correctness and Evidence. Use for analysis bottlenecks and cache or query design; CI/test-suite optimization belongs to rea-ci-tests-audit. | Stars: 472k | Updated | Claude CodeCodex | Verified | Install |
| rea-ci-tests-auditmorluto/reaAudit REA CI and test suites for correctness, fixture drift, false passes, and measured performance; use for suite or workflow investigations and efficiency changes, rather than ordinary feature testing or a single test failure. | Stars: 472k | Updated | Claude CodeCodex | Verified | Install |
| rea-tool-designmorluto/reaAudit, investigate, design, or change REA investigation tools, CLI/MCP contracts, provider capabilities, and Evidence semantics. | Stars: 472k | Updated | Claude CodeCodex | Verified | Install |
| ponytail-auditDietrichGebert/ponytailQuality audit of the whole repo: bugs, security, real load, missing tests, speed, and what to delete. Most important first. | Stars: 194k | Updated | Claude CodeCodex | Verified | Install |