- Home
- Skills
- Code Quality & Review
- tech-debt-scanner
More screenshots
Works with the AI Agents you already use
Tech Debt Scanner
Stop shipping debt with every feature — get a severity-rated inventory of every code smell, stale dependency, and untested path in minutes
Free
tech-debt-scanner
Example session with this skill installed
"Scan the billing service at ./services/billing/ for tech debt. It's a Python monolith, 40k LOC."
- Read your context and instructions
- Compiled the tech-debt-scanner
# Tech Debt Report — billing-service
Date: 2026-09-21 | Files Scanned: 127 | LOC: 40,200
SUMMARY
═══════
Critical: 3 | High: 4 | Medium: 6 | Low: 2
CRITICAL ITEMS
══════════════
File Line Category Finding First Action
──────────────────── ──── ──────────── ────────────────────── ────────────────────────
billing_service.py 847 Comment Debt FIXME in payment Read surrounding
reconciliation — function; extract error
security-adjacent, no handling to dedicated
error handling module
payment_processor.py 312 Test Debt Skipped test on Restore test or document
charge-refund flow — why it was disabled
highest-churn file (87
commits)
models/invoice.py 1 Structural 980 lines, 6-level Extract nested
nesting, 12 validation logic into
commits/month validators/
HIGH ITEMS
══════════
────────────────── ──── ────────── ──── ────────────
billing_service.py 234 Comment De
Connects securely to your tools. The creator never sees your data.
About this skill
### The Problem Your codebase has accumulated invisible debt: TODOs in auth code from 18 months ago, a 980-line billing module that changes every sprint, test-to-source ratio of 0.3 on your most critical paths, and lockfiles nobody has touched since the last framework upgrade. You know it's there — you feel it every time you open a file and spend ten minutes finding where to make the change. But you can't quantify it, you can't prioritize it, and you definitely can't get a sprint backlog out of "yeah, the codebase is messy." Without a structured scan, tech debt stays in the realm of vibes. Teams argue about what to fix, fix the wrong things, or (most commonly) fix nothing because the problem feels too big to start. ### What You Get - **Automated five-category debt scan** — comment debt (TODO/FIXME/HACK markers), structural debt (long functions, deep nesting, oversized files), dependency debt (stale lockfiles, EOL packages), test debt (missing tests, skipped tests, low ratios), documentation debt (undocumented APIs, orphaned docs) - **Severity classification with location multipliers** — a FIXME in auth code is Critical; one in a demo script is Low. Every finding carries file path, line number, risk statement, and a first remediation step - **Churn-weighted prioritization** — debt × change frequency compounds risk. A 600-line file touched 87 times is a higher-priority refactor than a 900-line file touched twice - **Skip-test detection** — finds every `it.skip`, `describe.skip`, `@pytest.mark.skip`, and `@Ignore` annotation, each needing a reason or restoration - **Anti-false-positive filtering** — generated code, vendored dependencies, TODOs with owner + ticket numbers, and documented workarounds are suppressed from findings - **Trend-tracking baseline** — markdown reports saved in a diffable format so re-scans after remediation sprints show measurable improvement ### Why This Beats Prompting It Yourself A raw `grep -rn "TODO|FIXME" .` gives you a flat list with zero context — every marker treated equally, no severity, no location intelligence, no churn data. The scanner applies a five-tier location multiplier (auth/payments/crypto → highest; scripts/demos → lowest) that transforms a 400-item grep dump into 15 actionable findings organized by sprint timeline. You get judgment, not just counts. ### Example Session You: "Scan the billing service at `./services/billing/` for tech debt. It's a Python monolith, 40k LOC." Agent: ``` # Tech Debt Report — billing-service Date: 2026-09-21 | Files Scanned: 127 | LOC: 40,200 SUMMARY ═══════ Critical: 3 | High: 4 | Medium: 6 | Low: 2 CRITICAL ITEMS ══════════════ File Line Category Finding First Action ──────────────────── ──── ──────────── ────────────────────── ──────────────────────── billing_service.py 847 Comment Debt FIXME in payment Read surrounding reconciliation — function; extract error security-adjacent, no handling to dedicated error handling module payment_processor.py 312 Test Debt Skipped test on Restore test or document charge-refund flow — why it was disabled highest-churn file (87 commits) models/invoice.py 1 Structural 980 lines, 6-level Extract nested nesting, 12 validation logic into commits/month validators/ HIGH ITEMS ══════════ ────────────────── ──── ────────── ──── ──────────── billing_service.py 234 Comment De ``` ### Use Cases - **Pre-release health gate** — scan before cutting a release to identify blockers and hidden risks - **Sprint planning input** — feed severity-rated findings directly into backlog grooming with estimated remediation effort - **New codebase onboarding** — generate a health map before touching unfamiliar code - **Refactor scoping** — identify which modules have the highest debt × churn compound risk - **Trend tracking** — baseline now, re-sprint, re-scan; the delta is what motivates remediation investment ### Known Limitations The scanner uses regex-based detection, not full AST parsing, so deeply nested or non-standard patterns may require manual review. Dependency staleness and churn analysis require a git repository — results outside git are limited to comment and structural debt only. ### Upgrade to Pro Free measures the debt. **Tech Debt Scanner Pro ($5)** manages it: debt-vs-churn heatmaps, automated cleanup tickets, architectural-drift tracking, and CI debt budgets. Upgrade when reduction has to survive a sprint plan — [Pro version](https://www.agensi.io/skills/tech-debt-scanner-pro).
The scanner uses regex-based detection, not full AST parsing, so deeply nested or non-standard patterns may require manual review. Dependency staleness and churn analysis require a git repository — results outside git are limited to comment and structural debt only.
How to use
- 1
Connect your agent to the Agensi MCP
Copy the prompt and paste it into Claude, ChatGPT, Cursor or any agent that supports MCP. It sets up the connection for you. Manual setup
- 2
Sign in to Agensi
Your agent shows a sign-in link. Sign in and click Allow.
- 3
Ask for what you need
Your agent loads this skill and uses it on your task. It always asks before any payment.
Reviews
No reviews yet
Be one of the first to try it. Every listed skill goes through the trust checks below.
Security scanned
Checked by our 8-point scan before listing
12 uses
Times used to date
Free forever
No account required to browse
Trust & safety
Security scan
Security scan passed 5 months ago
- Free with an Agensi account
Requires
Permissions
Allowed Hosts