360 Secure SkillHub
360 安全技能中心

Agentic Proposal Skill

为 CompleteTech LLC 代理式开发业务生成买方可见的商业文档:先判定商业阶段、选用提案/SOW 模板,再用经核实的事实填充占位符,最后用 Python 脚本渲染为品牌化 PDF(可选 PNG 预览)。文档含假设、排除项、验收标准与风险,缺失事实用 TBD 而不虚构。

completetech
completetech
热度 24.08
最近更新 2026-05-26 18:13:32
分类:营销依赖:代码执行 / 读写文件应用场景:文案创作 / 模板 / 报告撰写

安装技能

复制下面的话术发给你的 Agent,它会通过 secure-skills 检索并安装该技能。

执行命令安装此 skill: npx secure-skills-cli add agentic-proposal-skill
Metadata
name
agentic-proposal-skill
description
Create CompleteTech LLC agentic development proposals, statements of work, discovery recaps, pilot recommendations, evaluation plans, risk/control plans, implementation roadmaps, acceptance criteria, change orders, retainer proposals, and procurement-ready scope summaries. Use when Codex needs to bridge outreach emails and signed contracts/invoices by drafting buyer-facing proposal or SOW-style documents for bounded agentic workflow services.
version
1.0.2
metadata
{"openclaw":{"skillKey":"agentic-proposal-skill","homepage":"https://github.com/CompleteTech-LLC/agentic-proposal-skill","requires":{"bins":["python3"]},"install":[{"kind":"uv","package":"reportlab>=4.0"},{"kind":"uv","package":"pyyaml>=6.0"}]}}

Agentic Proposal Skill

Purpose

Create proposal and SOW-style sales documents for CompleteTech LLC agentic development engagements, from discovery recap through buyer approval.

System Boundary

This skill owns buyer-facing commercial scope before signature: proposals, SOW attachments, approval summaries, change-order proposals, and pilot recommendations. Use agentic-discovery-skill for upstream fact collection, agentic-contract-skill for agreement PDFs and legal-template artifacts, agentic-invoice-skill for payment requests, agentic-delivery-skill after approval, and agentic-email-skill only for the covering message that sends or follows up on the proposal.

Core Workflow

  1. Identify the commercial stage: post-discovery recap, pilot proposal, full proposal, SOW, procurement summary, change order, expansion, retainer, or evaluation/risk plan.
  2. Gather only verified facts: client, workflow, pain, systems, timeline, budget model, stakeholders, decision criteria, constraints, approval rules, and desired next step.
  3. Use references/use-case-decision-table.md to choose the right proposal type.
  4. Use references/proposal-positioning.md for brand language, scope boundaries, proof rules, and guardrails.
  5. Use references/proposal-catalog.md for near-exhaustive templates.
  6. Draft in a practical, bounded, low-hype voice. Do not fabricate case studies, savings metrics, regulated-use assurances, legal claims, or client facts.

Proposal Selection Guide

  • Post-discovery summary: use discovery-recap.
  • Short buyer-friendly pilot pitch: use one-page-pilot-proposal.
  • Full services proposal: use full-agentic-development-proposal.
  • Contract-ready scope attachment: use statement-of-work.
  • Fixed-price first engagement: use fixed-fee-pilot-proposal.
  • Hourly or flexible scope: use time-and-materials-proposal.
  • Ongoing support/monitoring: use monthly-retainer-proposal.
  • Advice without build: use advisory-only-proposal.
  • Workflow assessment before implementation: use workflow-assessment-proposal.
  • Multi-phase delivery plan: use implementation-roadmap.
  • Testing and acceptance focus: use evaluation-plan.
  • Governance or safety concern: use risk-control-plan.
  • Approval architecture: use human-in-the-loop-approval-model.
  • Added scope after agreement: use change-order-proposal.
  • Second workflow after pilot: use expansion-second-workflow-proposal.
  • Procurement or internal approval: use procurement-ready-scope-summary.
  • Executive decision maker: use executive-summary.
  • Buyer needs clear done criteria: use acceptance-criteria.
  • Scope boundaries unclear: use assumptions-and-exclusions.
  • Closeout and operations planning: use support-and-handoff-plan.
  • After discovery when a recommendation is needed: use post-discovery-recommended-pilot-memo.
  • Technical audience: use technical-architecture-proposal.
  • Data/tool access concerns: use data-and-tooling-readiness-proposal.
  • Training and enablement: use training-enablement-proposal.
  • Proof of concept only: use proof-of-concept-proposal.
  • Production rollout after pilot: use production-rollout-proposal.

When several templates fit, choose by buyer decision need first, then stage, then document length. If the buyer has not agreed on a workflow and success criteria, do not jump to a full SOW; use discovery recap, assessment, or pilot recommendation first.

Quality Rules

  • Frame agentic development as practical workflow implementation: discovery, tool routing, retrieval, approval gates, evaluation examples, logs, monitoring, documentation, support, and handoff.
  • Keep humans in the loop for external communications, production changes, purchases, data export, and material business decisions.
  • Include assumptions, exclusions, acceptance criteria, risks, and dependencies when the document is used for buying approval.
  • Make scope concrete enough to become a contract or invoice input.
  • Use TBD or ask for facts rather than inventing client details, legal terms, metrics, or proof.

Resource Guide

  • references/proposal-positioning.md: load for CompleteTech LLC offer framing, language, proof rules, and boundaries.
  • references/use-case-decision-table.md: load when deciding which document type to use.
  • references/proposal-lifecycle.md: load for end-to-end flow from discovery through contract/invoice handoff.
  • references/proposal-catalog.md: load for the near-exhaustive proposal/SOW template library.
  • references/template-index.json: machine-readable template metadata used by the renderer.
  • scripts/render_proposal.py: list proposal templates or render a draft with placeholders.

Renderer

python3 scripts/render_proposal.py --list
python3 scripts/render_proposal.py --stage pilot --list
python3 scripts/render_proposal.py --template one-page-pilot-proposal --var client_name=Acme --var workflow="support triage"

Rendered templates are drafts. Replace placeholders with verified facts and refine the narrative for the buyer.

Rendering to a Branded PDF

Artifacts from this skill are delivered as branded CompleteTech LLC PDF documents, not raw Markdown. The renderer emits the PDF (and prints the Markdown) in one command, using the same reportlab branding engine as the contract skill:

pip install -r requirements.txt
python3 scripts/render_proposal.py --template one-page-pilot-proposal \
  --out artifact.pdf --png artifact.png \
  --title "Support Email Triage Agent — Pilot Proposal" --doc-type "PROPOSAL / STATEMENT OF WORK" \
  --subtitle "Prepared for <b>Northwind Trading Co.</b>" --meta "PROPOSAL NO.=PRO-2026-0188" --meta "DATE=2026-05-20" \
  --var client_name="Client Name" --var workflow="support triage"
  • --no-pdf emits Markdown only (the original behavior); --no-cover drops the cover page.
  • Already drafted the Markdown yourself? Render it directly: python3 scripts/render_pdf.py --markdown artifact.md --out artifact.pdf --logo assets/logo.png --title "...".
  • The PDF supports a Markdown subset: #/##/### headings, paragraphs, - bullets, tables, > callouts, **bold**, and [PAGE_BREAK]. PDF requires reportlab; the optional --png preview requires pypdfium2 and pillow. See assets/examples/ for a rendered example.

Network Boundary

This skill is local-only. It does not include outbound network helpers, callbacks, or any helper that posts proposal run metadata to an external service.