Preface

Teams involved in government and enterprise bidding often face a common pain point: manual, fragmented workflows spread across Excel, search engines, and documents for drafting bids, interpreting scoring criteria, and finding candidates. Another pain point is data-related: whether external bidding data services can be integrated into a private environment, whether data egress occurs after integration, and whether it can be verified—these issues are often a “confused account” (lack clarity).

What is introduced below is a Skill: the Private Deployment version of the AI Bid System SaaS. It packages processes such as “finding candidates, interpreting scoring and rejection criteria, drafting technical and commercial bids, and auditing against the tender documents” into a capability callable by an Agent. One point must be clarified first: it is a Skill, not a pre-deployed SaaS system. Intranet operation, push notifications, knowledge bases, and security isolation require separate construction and verification.

What is this

One-sentence positioning: An Agent Skill designed for bidding assistance scenarios, covering four stages: candidate search, interpretation of scoring and rejection criteria, bid drafting, and auditing.

The publisher is “金里程信息” (Jinlicheng Information). The source is tagged as enterprise, the category is business-ops, the current version is 1.0.0, and the coordinate is @org-unhar74t/ai-bid-saas-private-deployment.

Core Functions

According to the official description, this Skill provides four types of capabilities:

  1. Find Candidates: Assists in searching for bidder candidates that meet specific conditions.
  2. Interpret Scoring & Rejection Criteria: Helps understand the scoring methods and bid rejection (veto) clauses in tender documents.
  3. Draft Technical & Commercial Bids: Generates initial drafts for both technical and commercial bids.
  4. Audit against Tender Documents: Checks the drafted bid against the tender document item by item.

The data source is the Shibo (Shibo) Bid Data API, covering announcements, winning bids, procurement intentions, and enterprise monitoring data. The actual coverage and permissions are subject to the service provider. The Skill does not implement automatic monitoring—it requires active querying and will not proactively watch for updates on specific projects.

Queries default to nationwide scope, the last 30 calendar days, and the top 10 results; analysis defaults to the last 6 months, and the output discloses the sample scope and limitations. This implies that the results are sampling-based, not full statistical counts, and citations should include these limitations.

Installation and Activation

The official page does not provide specific installation commands or configuration scripts. The verifiable prerequisites for activation are as follows:

Tags: requires_api_key=true
Data Source: Shibo Bid Data API

You need to prepare a Shibo API key before use; otherwise, functions involving data queries will not run. The Skill is based on the generic SKILL.md format. It can be enabled by loading the Skill directory in Agents that support this format (such as Cursor, Codex CLI, Claude Code). Specific installation commands, environment variable names, and field configurations should follow the official page; they are not speculated here.

Typical Usage

The materials do not provide reproducible complete command-line examples. The invocation methods divided by function are given below; specific parameters are subject to the official documentation:

Find Candidates: Search for bidding candidates based on conditions
Interpret Scoring: Input scoring methods and rejection items, output interpretation
Draft Bid: Generate initial drafts for technical and commercial bids respectively
Audit Bid: Audit bid content against tender document items one by one

When executing queries and analyses, pay attention to the default parameters: nationwide, last 30 calendar days, top 10 results; analysis window of the last 6 months. If a longer period or more entries are required, confirm whether the service permissions support it.

Applicable Scenarios & Precautions

Who is it for: Government/enterprise bidding teams, practitioners needing to batch draft or audit bids, and developers looking to integrate bidding assistance capabilities into self-built Agents.

Limitations and Pitfalls:

  1. A Skill is not a deployed SaaS. Intranet operation, message pushing, knowledge base management, and security isolation must be separately built and verified; do not assume they exist by default.
  2. External API does not mean data stays offline. After connecting to the Shibo API, requests will pass through external services. Data flow depends on your deployment configuration and network policies. Confirm compliance requirements before use.
  3. Data coverage and permissions are subject to the service. Automatic monitoring is not implemented.
  4. Query and analysis results have default sample limitations (top 10 results from the last 30 calendar days, 6-month analysis window). Include sample descriptions when citing data.
  5. Scoring, rejection interpretation, and bid drafts require human review. They cannot be used directly as final bidding decisions.

Conclusion

The value of this Skill lies in embedding four common bidding assistance stages into an Agent. Data queries have clear default sample boundaries, and limitations are clearly stated in the introduction. The prerequisite for application is that you have confirmed the data compliance of the external API, and you are prepared to perform separate verification for intranet deployment, push notifications, and the knowledge base.

Official URL: https://api.skillhub.cn/org-unhar74t/ai-bid-saas-private-deployment