Deal Review logo

Deal Review

Community
davekilleen
deal-review

Review active deals from canonical activity evidence and surface unknowns

Overview

Publisherdavekilleen
RepositoryDex
Skill namedeal-review
Stars
481
Forks
130
Bundled files
Instructions only
Links
  • Markdown instructions

    A SKILL.md file the model loads on demand, so it only costs tokens when a request actually matches.

  • Works with any LLM

    AI skills are plain Markdown, not provider-specific code, so this works with GPT, Claude, Gemini, Grok, or a local model.

  • Self-contained

    Everything the model needs lives in the instructions — no extra files to sync.

  • Open source

    Published by davekilleen on GitHub. Read the source before you install it.

Installation

Install the Deal Review AI skill in TypingMind to use it with any LLM, or drop it into another agent that reads SKILL.md.

1

Install in TypingMind

TypingMind installs a skill straight from its GitHub folder — it reads SKILL.md, bundles the resource files, and stores the result locally.

  1. Open the app and go to Plugins → Skills.
  2. Choose "Install from GitHub".
  3. Paste the skill folder URL below and confirm.
  4. Enable the skill in any chat where you want it available.
Plugins → Skills → Add skill → From GitHub URL, then paste the folder URL and press Continue.
2

Install in another agent

Any agent that reads the Agent Skills format can use this skill — copy the folder into that agent's skills directory.

Claude Code — .claude/skills
git clone --depth 1 https://github.com/davekilleen/Dex.git /tmp/Dex
mkdir -p .claude/skills
cp -r /tmp/Dex/packages/dex-agent-plugin/skills/_available/sales/deal-review .claude/skills/deal-review
Restart Claude Code after copying so it picks up the new skill.

Use it in TypingMind

Enable Deal Review in any TypingMind chat and the model takes it from there. Its name and description sit in the system prompt, and the moment a request matches, the model loads the full instructions itself — you never invoke it by hand, and it costs no tokens until it is actually used.

The model loads Deal Review on its own as soon as a request matches it.

Works with any AI model

AI skills are plain Markdown instructions rather than provider-specific code, so Deal Review is not tied to the model it was written for. Install it once in TypingMind and use it with GPT-5, Claude, Gemini, Grok, DeepSeek, Mistral, Llama, or a local model you run yourself — all on your own API keys.

  • Loaded only when it is needed

    The system prompt carries just the name and description. The instructions are fetched on the first matching request, so an idle skill costs nothing.

  • Switch models mid-chat

    Because the skill is instructions rather than code, changing model does not break it — the next model reads the same SKILL.md.

Skill instructions

This is the SKILL.md content the model loads. Read it before installing — a skill is instructions your model will follow.

Purpose

Create a complete, dated review of the requested deal cohort. Surface evidence-backed risks, missing next steps, deadlines, and unchecked deals while keeping unknowns visible.

Usage

  • /deal-review — review the confirmed active-deal cohort
  • /deal-review [stage] — filter through the configured stage map
  • /deal-review [period] — filter by sourced close dates in a confirmed period

Evidence, authority, and recovery

Set the review as-of timestamp and timezone first. Use the canonical activity date: the actual timestamp on a meeting, call, email, CRM event, or other authoritative record. Store its source ID/path and source date. A file modified date is only a weak discovery clue and must not become an activity.

  • Keep an explicit Unchecked deals section for every discovered record that is unreadable, duplicated, contradictory, or missing a required field. Never silently drop an unchecked deal or call it healthy.
  • Unknown value must be excluded from value totals and amount-share denominators; never replace it with zero. An explicit zero remains eligible and is labelled.
  • Disclose denominator coverage for every rate, count, or percentage: discovered, checked, eligible, and excluded rows.
  • Keywords and silence are leads, not facts. Do not infer ghosting, churn, an absent buyer, or a blocker without dated source evidence.
  • Never invent absent facts, dates, values, next steps, commitments, risks, or confidence.
  • Keep the review read-only. Before any requested mutation, preview the exact target and payload/diff, require confirmation from the authorized human, and read back the result. If the write or read-back fails, report possible partial state and stop.

Method

1. Confirm the cohort

Confirm the authoritative deal source, requested period, stage filter, currency/units, and which states count as active. If active-state policy is absent, show discovered states and ask rather than silently selecting a cohort.

Search configured deal locations and linked account/person records. Deduplicate only on a stable deal ID or a human-confirmed mapping; similar names are not proof of identity.

2. Build the deal ledger

For every discovered deal capture:

FieldHandling
Deal ID and accountstable source identity
Valueamount, currency, source/date, or Unknown
Stageconfigured value and source/date, or Unknown
Last activitycanonical event, event date, and source
Next stepexact text, owner, due date, and source
Close datedate and source, or Unknown
Stakeholderssourced roles; do not infer missing roles
Risk evidenceexact dated observation, not a keyword alone

Show contradictory records side by side and leave the field unresolved until a human with authority decides.

3. Assess freshness from policy

Calculate elapsed time only from a canonical activity date. Apply Fresh, Aging, Stale, or any urgency label only when the user supplies or confirms a dated freshness policy for this cohort. Record that policy source in the report.

If no policy exists:

  • report Last canonical activity: [date / Unknown];
  • report factual elapsed time where possible;
  • set Freshness assessment: Unknown — no confirmed policy;
  • offer a question, not a judgment.

4. Assess next steps, deadlines, and risk

A next step is complete only when the confirmed policy's required fields are present. Without a policy, show the exact text and missing owner/date rather than assigning a red/amber/green label.

For a requested deadline window, use sourced close and due dates. Do not assume a timeline is realistic from stage alone.

Treat terms such as “blocked,” “competitor,” or “waiting” as search leads. Cite the underlying sentence/event, distinguish fact from hypothesis, and show contradictions. A risk classification needs either explicit source evidence or a configured rule.

5. Reconcile and verify

  • Reconcile checked plus unchecked counts to the discovered cohort.
  • Reconcile stage counts to eligible checked deals.
  • Reconcile known-value subtotals to the known-value total.
  • List all unknown-value exclusions.
  • Verify each recommendation points to an observed fact or a named unknown.

Output contract

markdown
# Deal review

**As of:** [timestamp and timezone]
**Cohort definition:** [source/policy]
**Discovered / checked / unchecked:** [N / n / u]
**Known-value coverage:** [n/N; currency and exclusions]

## Needs attention
### [Deal]
- Stage: [value + source/date or Unknown]
- Last canonical activity: [event/date/source or Unknown]
- Freshness: [policy-backed label or Unknown]
- Next step: [exact sourced text or Unknown]
- Risk evidence: [fact / hypothesis / contradiction]
- Suggested question or action: [read-only recommendation]

## On track
[Include only when the configured policy is satisfied; cite the policy and evidence.]

## Unchecked deals
| Deal | Reason | Last successful source read |
|---|---|---|

## Pipeline distribution
[Counts and known values with explicit denominators and exclusions.]

## Unknowns and contradictions
- [Both sources/dates and what needs human resolution]

## Recommended actions
1. [Action tied to evidence; no write performed]

Placeholders define shape only. Never populate them from generic examples.

Controlled follow-up

For a requested deal or task update:

  1. read the authoritative current target;
  2. preview the exact before/after values or payload;
  3. identify related records that will not be changed;
  4. require explicit human confirmation;
  5. write only the confirmed target;
  6. read back and compare every field with the preview.

On failure, preserve prior content where possible, re-read current state, disclose any partial result, and require a fresh preview before retrying.

Frequently asked questions

What does the Deal Review AI skill do?

Review active deals from canonical activity evidence and surface unknowns

Why use Deal Review on TypingMind?

Because you install it once and use it with any model. Deal Review is plain Markdown rather than provider-specific code, so the same skill runs on GPT-5, Claude, Gemini, Grok, or a local model — and you can switch model mid-chat without it breaking. TypingMind runs on your own API keys, so you pay providers directly instead of a per-seat subscription, and your skills and chats stay in your own storage.

How do I install Deal Review in TypingMind?

Open Plugins → Skills → Install from GitHub in TypingMind and paste https://github.com/davekilleen/Dex/tree/main/packages/dex-agent-plugin/skills/_available/sales/deal-review. TypingMind reads its SKILL.md and installs it as a skill you can enable per chat.

Which AI models can use Deal Review?

Any model you connect in TypingMind. AI skills are plain Markdown instructions rather than provider-specific code, so GPT, Claude, Gemini, Grok, and local models can all load this skill when a request matches it.

How many AI models can I use with Deal Review?

As many as you like. As long as a model supports skills, you can use Deal Review with it — GPT, Claude, Gemini, Grok, DeepSeek, Mistral, Llama and more — all on TypingMind with your own API keys.

Is the Deal Review AI skill free?

It is published on GitHub by davekilleen. Check the repository for licensing terms. You only pay your own AI provider for the tokens you use.

What are AI skills?

An AI skill is a reusable instruction bundle that teaches an AI model how to do one specific task. It follows the open Agent Skills format: a SKILL.md file with a name and description, plus any scripts, templates or reference files the model may need. The model reads the instructions only when your request matches the skill, so an installed skill costs nothing until it is used.

How are AI skills different from plugins or MCP servers?

A plugin or MCP server gives a model new tools to call — code that runs somewhere and returns a result. An AI skill gives the model knowledge and process instead: how to approach a task, which steps to follow, what good output looks like. Skills are plain Markdown, so they need no server, no API key and no runtime, and they work with any model.

View all

Set up your own AI workspace now

Get notified about new features and future giveaways by subscribing to our newsletter 👇