Score SME financing candidates with a deterministic rule-based rubric.
npx skills add https://github.com/mr-kelly/skills --skill kelly-deal-scorer
Please help me install this Agent Skill. Skill: kelly-deal-scorer Source: https://github.com/mr-kelly/skills/tree/main/skills/kelly-deal-scorer Command: npx skills add https://github.com/mr-kelly/skills --skill kelly-deal-scorer
This skill removes the ambiguity and black-box nature of deal underwriting by providing a fully transparent, rule-based scoring system for SME financing candidates.
Invoke the deal scoring desk by running the local launcher script and opening the provided URL in your browser to begin reviewing the candidate queue.
High-intent search queries and answers about installing and using this skill.
You score SME financing deals by applying a deterministic, arithmetic-based rubric that computes a 0-100 composite score with a factor-based breakdown. This ensures every underwriting decision is fully transparent, repeatable, and auditable.
Deterministic underwriting scoring calculates a 0-100 composite score for SME candidates using fixed arithmetic rules rather than opaque evaluations. It provides factor-based score breakdowns and suggested revenue-share ranges for auditable financing decisions.
You set up a local review queue dashboard by running a local launcher script that starts a Hono server environment. This provides a zero-build vanilla frontend in your browser to track candidate status, view revenue history, and record human decisions.
Yes, you can rank SME financing candidates by revenue stability, growth trends, and risk tiers using a rule-based scoring system. This approach eliminates the black-box nature of LLM evaluations, ensuring fully transparent and auditable underwriting decisions.
Yes, the deal scoring desk requires a local Hono server environment to manage file-backed state and provide the zero-build vanilla frontend for human review. You need Node server dependencies installed to run the local launcher script.
The rule-based rubric requires local candidate batches to operate and does not rely on adaptive LLM-based evaluations. Scoring is strictly deterministic, meaning the customizable arithmetic rubric must be manually adjusted to change ranking factors or revenue-share ranges.
mr-kellyRun local LLM-agent regression reviews against a baseline with mock test cases.
jikig-aiForecast company-level revenue from aggregate monthly and quarterly data.
bitwize-music-studioGenerate an album progress dashboard with per-phase status and track blockers.
mr-kellyModel revenue-based financing deals and calculate projected cash flows and repayment metrics.
