Skip to content
Skills Library
AI & agents · Review & audit

tool-stack-audit

/tool-stack-audit

Use when a founder wants to evaluate their current AI tools — identify overlap, waste, gaps, and whether tools match actual use cases

Licence
MIT
Last repository change
03.10.2026

Assessment and security

100/ 100Class: good
Format25 / 25
Frontmatter with name and description, valid name, description up to 1,024 characters, name matching the folder.
Discoverability25 / 25
Description explains when to use the skill, has a reasonable length and uses third person.
Content35 / 35
At least 150 words and at most 500 lines, headings, step lists, examples or code.
Hygiene15 / 15
No placeholders such as TODO, no Windows paths, with a licence.
Riskno signals detected in the scan
Files in this package
1
Scripts
0
HTML comments
0
URLs in SKILL.md
1
Deprecation

No signals detected. Repository changed on 03.10.2026, without references to retired models or APIs.

Licence

MIT, open: use and modify while retaining author attribution.

What it does and when to use it

Use when a founder wants to evaluate their current AI tools — identify overlap, waste, gaps, and whether tools match actual use cases

Original author description

Use when a founder wants to evaluate their current AI tools — identify overlap, waste, gaps, and whether tools match actual use cases

How it works

  1. Purpose
  2. Context Intake
  3. Process
  4. Output
  5. Anti-Patterns
  6. Department Profiles
  7. References
Original SKILL.md headings · 145 lines, 1,005 words
SKILL.md preview
---
user-invocable: true
name: tool-stack-audit
description: Use when a founder wants to evaluate their current AI tools — identify overlap, waste, gaps, and whether tools match actual use cases
---

# Tool Stack Audit

## Purpose

Structured audit of a team's AI tools: what they have, what's used, what's wasted, and what's missing. Produces a clear inventory with cost analysis. This is an audit, not a shopping list — it evaluates what exists before recommending changes.

**Core principle:** Audit what you have before buying what you don't. Most teams have more tools than they use and more overlap than they realize.

## Context Intake

For `Department:` and `Currency:`, use the first available source: the current fluency scorecard → `adoption.local.md` (the department this run covers; by default the one marked `(primary)` — see CLAUDE.md Local Configuration) → ask the leader (currency defaults to USD).

## Process

<HARD-GATE>
1. Audit existing tools first. Do NOT recommend new tools until the current stack is assessed.
2. Use actual usage data, not what tools are "supposed to" do.
3. Categorize every tool: keep, consolidate, or cut. No tool stays uncategorized.
4. Flag shadow IT (individual subscriptions) as a cost and governance risk.
5. Do NOT recommend specific vendor products. Recommend capabilities the team needs.
</HARD-GATE>

### Required Inputs

- **Tool list:** Every AI tool the team pays for (team licenses and individual subscriptions)
- **Per tool:** Cost, number of seats, who uses it, what for, how often
- **Overlap:** Any tools that do similar things
- **Gaps:** Tasks where people want AI help but don't have a tool

### Audit Criteria

Evaluate each tool against four questions:

| Question | What you're looking for |
|----------|------------------------|
| **Is it used?** | Weekly active users vs. seats paid for |
| **Is it the right tool for the job?** | Does it match the actual use case, or was it bought for a different purpose? |
| **Does it overlap with another tool?** | Two tools doing the same thing = one should go |
| **Is it governed?** | Does the company control the account, or is it an individual subscription? |

## Output

Produce the audit in this exact format. **Currency note:** Substitute the `[CCY]` placeholder below with the currency symbol from the fluency-assessment scorecard (`$` for USD, `€` for EUR, `£` for GBP, etc.). If no scorecard is available, ask the founder which currency they report in.

```
## AI Tool Stack Audit
**Company:** [name] | **Team:** [size] | **Date:** [date]
**Currency:** [USD / EUR / GBP / other — from scorecard]

### Current Stack

| Tool | Cost/Month | Seats | Weekly Users | Utilization | Primary Use |
|------|:---------:|:-----:|:-----------:|:-----------:|-------------|
| [Tool 1] | [CCY][X] | [X] | [X] | [X]% | [use case] |
| [Tool 2] | [CCY][X] | [X] | [X] | [X]% | [use case] |
| Individual subs (estimated) | [CCY][X] | [X] | [X] | — | Various |
| **Total** | **[CCY][X]** | | | | |

### Assessment

| Tool | Verdict | Reason |
|------|---------|--------|
| [Tool 1] | Keep / Consolidate / Cut | [One sentence] |
| [Tool 2] | Keep / Consolidate / Cut | [One sentence] |

### Overlap Found
[List any tools with overlapping capabilities and which one to keep, with reasoning.]

### Waste Identified
- Unused seats: [X] seats across [X] tools = [CCY][X]/month wasted
- Redundant tools: [list] = [CCY][X]/month overlap
- **Total recoverable spend: [CCY][X]/month ([CCY][X]/year)**

### Gaps
[Capabilities the team needs but doesn't have — described as needs, not product recommendations.
E.g., "No AI assistance for code review" (engineering) / "No AI meeting summaries that update the CRM" (sales) / "No AI drafting for recurring documents" (generic) — not "Should buy CodeRabbit / Gong / ChatGPT."]

### Shadow IT Risk
[Individual subscriptions outside company control: estimated count, estimated cost, governance concern.]

### Summary
[2-3 sentences: total spend, utilization rate, main action — keep/consolidate/cut.]

---
*Generated by [AI Adoption Playbook](https://github.com/adimango/ai-adoption-playbook) v1.3.0 on [YYYY-MM-DD]*
```

## Anti-Patterns

### The Shopping List
**Symptom:** Audit becomes a recommendation to buy 3 new tools.
**Consequence:** Founder spends more before fixing what they have. Tool sprawl gets worse.
**Fix:** Audit first, fully. Only identify gaps as capabilities needed, not products to buy.

### Ignoring Shadow IT
**Symptom:** Audit only covers company-paid tools.
**Consequence:** Missing 20-40% of actual AI spend. No visibility into what team members are actually using.
**Fix:** Always ask about individual subscriptions and expense reports.

### Recommending Specific Vendors
**Symptom:** "You should switch to Cursor" or "Try Sourcegraph Cody."
**Consequence:** Recommendation ages badly (tools change fast), may not fit team's needs.
**Fix:** Describe the capability gap. Engineering example: "Your team needs AI-assisted code review that integrates with your PR workflow." Sales example: "Your team needs AI meeting summaries that auto-update your CRM." Let the founder evaluate vendors.

## Department Profiles

Use the relevant profile to recognize typical tools and capability gaps for the team being audited.

### Engineering

**Typical tools to ask about:** GitHub Copilot, Cursor, Sourcegraph Cody, Tabnine, CodeRabbit, Codium, Claude Code, ChatGPT (individual), Sweep, Codeium.

**Common capability gaps (express as needs, not products):**
- AI-assisted code review that integrates with the PR workflow
- Test generation tied to the existing test framework
- Automatic doc drafts from code changes
- AI debugging assist inside the IDE

### Sales

**Typical tools to ask about:** Gong, Chorus, Clari, Salesloft, Outreach, Apollo, Lavender, Regie, ChatGPT/Claude (individual), HubSpot AI, Salesforce Einstein.

**Common capability gaps (express as needs, not products):**
- Meeting summaries that auto-update CRM fields
- AI-drafted outreach personalized from CRM data
- Pipeline risk scoring that flags stalled deals
- Proposal/deck first-draft generation from RFP input

### Generic

**Typical tools to ask about:** ChatGPT, Claude, Microsoft 365 Copilot, Google Gemini for Workspace, Notion AI, Otter, Fireflies.

**Common capability gaps (express as needs, not products):**
- Meeting summarization with action items
- Draft generation for recurring documents
- Research synthesis from multiple sources
- Workflow automation for repetitive steps

## References

- `adoption-scorecard` — provides usage data that feeds the utilization column
- `roi-calculator` — uses cost data from this audit
- `first-use-case-picker` — gaps found here inform which use cases to pursue

Versions

The database has no other skills named tool-stack-audit.

Database snapshot: 13.11.2026 · package version 72e3a7a097ed (SHA-256 digest of the entire package)

Downloads in the last 30 days: 0 · ranking

Report a problem with this skill: harmful content, incorrect descriptions or an author’s removal request.

Similar skills

  • 91stack-ai-agent
    Conventions for the AI agent stack (MCP tools per task, eval cases per tool, offline evals in the gate). Use when adding an MCP tool, an eval case, a prompt or a model call in this repo.
  • 90budget-audit
    Audit AI spend — per-model cost breakdown, top consumers, waste identification, optimization levers. Use when asked "why is our AI bill so high", "audit LLM spend", or "where is our token waste".
  • 100keel-audit
    Operational efficiency audit — identify waste, redundancy, and friction across processes, tools, and team workflows. Use when asked to "audit our operations for waste", "where are we inefficient", "what tools are we paying for but not using", or "reduce operational overhead".
  • 100context-radar-audit
    Audit a project's Claude Code tool stack (MCP servers, plugins, hooks) against the curated recommendations at Context Radar (thoroc.github.io/context-radar): a 90-plus-tool, 20-layer comparison of context-reduction, memory, and code-nav tools for coding agents. Use when asked to…
  • 100plugin-audit
    Audit vercel-plugin performance on real-world projects. Extracts tool calls from Claude Code conversation logs, tests hook matching against actual inputs, identifies pattern coverage gaps, and checks plugin cache staleness. Use when asked to audit, test, or investigate plugin…
  • 100tool-design
    This skill should be used when the user asks to "design agent tools", "create tool descriptions", "reduce tool complexity", "implement MCP tools", or mentions tool consolidation, architectural reduction, tool naming conventions, or agent-tool interfaces.

Feedback and reports

Your message goes to Bartosz Sokoliński. An email address is needed only for a reply. See the privacy policy.