aiskillstore/marketplace / skills/1102tools-dev/igce-builder-cr
igce-builder-cr
VerifiedTrigger for: cost-reimbursement IGCE, CR cost estimate, CPFF, CPAF, CPIF, cost-plus estimate, BAA estimate, fixed-fee analysis, award-fee analysis, incentive-fee analysis, proposed CR rate validation, cost-pool buildup, share-ratio scenario, price-reasonableness memo, or fair-and-reasonable analysis. Build auditable CR estimates using BLS OEWS wages, fringe/overhead/G&A/FCCM pools, contract-type-specific fee, GSA CALC+ positioning, and GSA Per Diem travel. Do NOT use for FFP, LH/T&M, grants, or cooperative agreements. Requires the bls-oews, gsa-calc, and gsa-perdiem MCP servers.
| File | Kind | Size |
|---|---|---|
| SKILL.md | skill md | 16.4 KB |
| agents/openai.yaml | config | 285 B |
| references/data-source-operations.md | markdown | 13.4 KB |
| references/runtime-adaptation.md | markdown | 2.7 KB |
| references/validation-gates.md | markdown | 4.2 KB |
| references/workbook-specification.md | markdown | 6.9 KB |
| references/wrap-rate-presets.md | markdown | 4.0 KB |
| scripts/recompute_expected_values.py | script | 9.0 KB |
| scripts/validate_workbook.py | script | 17.7 KB |
| skill-report.json | config | 45.8 KB |
| test.md | markdown | 2.6 KB |
| testing.md | markdown | 21.6 KB |
references/validation-gates.md
View on GitHub# CR Workbook Validation Gates
Run all applicable gates after workbook creation and before delivery.
## Layer 1: formula-structure audit
The bundled validator checks:
- All seven required sheets exist.
- `IGCE Summary!B11` computes the month gap from `YYYY-MM` text using `VALUE`, `LEFT`, and `MID`, without `DATEDIF` or `YEAR`.
- `IGCE Summary!B12` computes the aging factor from escalation and month gap.
- Every Cost Buildup block follows the 23-row layout.
- Direct labor is the fifth row of each block and cross-sheet formulas do not use the aged-wage row by mistake.
- Fringe, overhead, G&A, FCCM, fee, and price formulas preserve the required bases.
- The workbook contains formulas and no formula-error tokens in formula text.
- Text beginning with formula-trigger characters is escaped.
- Optional custom formula assertions in the JSON sidecar pass.
## Layer 2: independent recomputation
Create a JSON sidecar from the raw inputs, not from workbook formula results. Minimum shape:
```json
{
"assumptions": {
"fee_type": "CPFF",
"primary_fee_rate": 0.08,
"fringe_rate": 0.32,
"overhead_rate": 0.80,
"ga_rate": 0.12,
"fccm_rate": 0.0,
"aging_factor": 1.025,
"productive_hours": 1880
},
"labor_lines": [
{
"name": "Research Scientist",
"annual_wage": 140000,
"fte": 3,
"months": 12,
"workbook_cost_rate_cell": "'Cost Buildup'!B17",
"workbook_price_rate_cell": "'Cost Buildup'!B21"
}
],
"non_labor_lines": [
{
"name": "Travel",
"amount": 12000,
"fee_bearing": false,
"workbook_total_cell": "'Travel Detail'!B14"
}
],
"workbook_fee_bearing_cost_cell": "'IGCE Summary'!B28",
"workbook_total_cost_cell": "'IGCE Summary'!B29",
"workbook_total_fee_cell": "'IGCE Summary'!B30",
"workbook_total_price_cell": "'IGCE Summary'!B31"
}
```
For CPAF, add `award_pool_rate` and `assumed_earned`. For CPIF, the primary recomputation verifies the target case; add formula assertions for overrun and underrun shares and bounds in `Scenario Analysis`.
If continuous coverage is present, put `annual_coverage_hours` on the relevant labor line. The recomputation rejects an FTE and productive-hours combination that does not reconcile within 0.5% or one hour.
Run:
```text
python scripts/recompute_expected_values.py validation-input.json
python scripts/validate_workbook.py output.xlsx --expected validation-input.json --engine auto
```
## Layer 3: real spreadsheet engine
With `--engine auto`, the validator looks for LibreOffice. When found, it performs a headless conversion cycle, reopens the calculated workbook with `data_only=True`, checks cached errors, and compares named workbook values against the independent Python result.
If formula execution cannot run, do not call the workbook fully recalculated. Use the exact disclosure required by the core skill.
## CR-specific manual gates
Confirm each item even when a script passes:
- Fee-bearing and non-fee-bearing amounts are visibly separate.
- CPFF form is Completion or Term and the rationale follows FAR 16.306(d).
- CPFF fee does not exceed the applicable ceiling.
- CPAF shows base-only, assumed-earned, and full-pool outcomes.
- CPIF uses separate contractor overrun and underrun shares and applies min/max bounds.
- Audited indirect rates are point estimates with date and authority, not scenario midpoints.
- CALC+ results are neutral positioning and distinguish estimated cost from cost plus fee.
- Day-trip M&IE is discounted once, not twice.
- BLS vintage is runtime-confirmed and aging is cell-referenced.
- No formula range contains text `TBD`.
## Fault injection
At minimum, test copies with these deliberate defects and confirm rejection:
1. Replace `keyword=` with `q=` in a recorded CALC+ query.
2. Change the month-gap formula to `DATEDIF` or `YEAR` on text.
3. Point a summary rate to the aged-annual-wage row instead of Direct Labor.
4. Apply fee to total estimated cost when a non-fee pass-through is present.
5. Remove the FCCM layer or apply it to the wrong base.
6. Change a CPIF underrun formula to use the overrun share.
7. Multiply 4.2 FTE by 1,880 hours against an 8,760-hour coverage requirement.
Document which faults were automated and which were manually inspected in `test.md`.
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 |