Skip to content

BdM-15/capture-insights / skills/proposal-generator

proposal-generator

Verified

Shipley-methodology federal proposal outline and section drafter. USE WHEN the user asks to draft a proposal volume, build an outline from the proposal_instruction ↔ evaluation_factor traceability (UCF Section L/M or equivalent for non-UCF — FAR 16 task orders, FOPRs, BPA calls, OTAs, agency-specific formats), generate a compliance matrix, write win themes, draft an executive summary, propose FAB (Feature → Advantage → Benefit) chains, identify discriminators, or 'respond to this RFP'. Pulls requirements, evaluation factors, instructions, customer priorities, and pain points from the active Theseus workspace KG and produces an evidence-cited draft. Also ships govcon HTML render templates under assets/ — hand the rendered content off to the `huashu-design` skill for PPTX / PDF / animation export. Format-agnostic — never assumes UCF section labels are present. DO NOT USE FOR clause compliance auditing only (use compliance-auditor) or extracting new entities (use govcon-ontology + the Theseus pipeline).

evals/evals.json

View on GitHub
{
  "skill_name": "proposal-generator",
  "evals": [
    {
      "id": 1,
      "prompt": "Generate the compliance matrix for this RFP, mapping every proposal instruction to the corresponding evaluation factor and underlying requirement.",
      "expected_output": "A JSON artifact at artifacts/proposal_draft.json whose compliance_matrix array contains one row per proposal_instruction entity in the workspace, each row populated with instruction_id, evaluation_id, requirement_ids, source_chunks, and a status of OK | PARTIAL | GAP. The final assistant message is a short cover note summarizing row counts.",
      "expectations": [
        "The skill called kg_entities at least once requesting types including 'proposal_instruction' and 'evaluation_factor'.",
        "The skill called kg_chunks at least once with a query mentioning 'instructions', 'evaluation', or 'section L'/'section M'.",
        "The skill called write_file with target path 'artifacts/proposal_draft.json'.",
        "The persisted JSON contains a top-level 'compliance_matrix' array.",
        "Every compliance_matrix row whose status is 'OK' carries at least one chunk_id in 'source_chunks'.",
        "No row has status 'GAP' attributable solely to a missing 'Section L' or 'Section M' label."
      ]
    },
    {
      "id": 2,
      "prompt": "Build win themes for this opportunity. Tie every theme to a customer_priority or pain_point in the workspace, and identify the discriminator and proof point.",
      "expected_output": "Themes array in artifacts/proposal_draft.json where each theme object has theme, discriminator, proof, hot_button_id (matching a customer_priority or pain_point entity), and source_chunks. The cover note in the final assistant message lists the theme count.",
      "expectations": [
        "The skill called kg_entities requesting 'customer_priority' and 'pain_point' (and optionally 'past_performance_reference').",
        "The skill read assets/theme_card.md via read_file before drafting themes.",
        "Every theme.hot_button_id maps to an entity name returned by kg_entities.",
        "No theme is phrased as a bare adjective-noun like 'Innovative Solutions' or 'World-Class Service'.",
        "The artifact's themes array is non-empty if customer_priority or pain_point entities exist in the workspace."
      ]
    },
    {
      "id": 3,
      "prompt": "Draft the executive summary for Volume I. It must address the customer mission, our top three discriminators, and call to action — maximum two pages.",
      "expected_output": "executive_summary_md field populated in artifacts/proposal_draft.json. The summary references at least three discriminators that also appear in fab_chains, cites chunk_ids for any direct quotes, and stays under roughly 2 pages (~1200 words).",
      "expectations": [
        "The skill executed the spine + themes + FAB workflow steps before drafting the executive summary (assistant turns show kg_entities and theme drafting before executive_summary_md is written).",
        "The skill read assets/executive_summary.md via read_file.",
        "executive_summary_md in the artifact is non-empty and under 1500 words.",
        "Every direct quote of RFP language in executive_summary_md is followed by an inline [chunk-xxxxxxxx] citation."
      ]
    },
    {
      "id": 4,
      "prompt": "Generate the compliance matrix for this RFP, then return the standard cover note.",
      "expected_output": "A short Markdown cover note that summarizes the artifact via the bullet shape defined in the Final Assistant Message section (matrix rows, themes, FAB chains, volumes, warnings, top-warnings list, sources cited). Cover notes must NOT restate facts about the workspace as standalone narrative claims; substantive evidence lives in artifacts/proposal_draft.json's source_chunks fields.",
      "expectations": [
        "tools/audit_skill_grounding.py --skill proposal-generator reports grounding_ratio is >= 0.85 (cover-note-style skill; the JSON artifact's source_chunks fields are the citation-of-record, the assistant message is a structural summary; 147.10 calibration credits structural envelope-field readouts and traceability-gap findings as exempt).",
        "tools/audit_skill_grounding.py reports claims_exempt > claims_unsourced (most cover-note prose should be recognized as structural / pointer / self-attestation, not unsourced claims).",
        "No regression below 0.50 across at least three consecutive runs of the same prompt."
      ]
    }
  ]
}
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: 471kUpdated Claude CodeCodexVerifiedInstall
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: 471kUpdated Claude CodeCodexVerifiedInstall
rea-tool-designmorluto/reaAudit, investigate, design, or change REA investigation tools, CLI/MCP contracts, provider capabilities, and Evidence semantics.Stars: 471kUpdated Claude CodeCodexVerifiedInstall
ponytail-auditDietrichGebert/ponytailQuality audit of the whole repo: bugs, security, real load, missing tests, speed, and what to delete. Most important first.Stars: 194kUpdated Claude CodeCodexVerifiedInstall