I Gave My AI Coding Agent 5 Skills and It Stopped Rewriting the Same Files

The Android AI Architect — Mon, 05 Oct 2026 09:22:38 +0000

Most AI coding setup advice stops at "install Claude Code." That's the easy part.

The hard part is what happens three weeks in: the agent keeps rewriting the same helper file, forgets the project conventions, writes code without tests, and you spend more time reverting than shipping.

Skills are just instructions your agent reads

An agent skill is a markdown file with a name, a description, and a procedure. No plugin system, no framework, no build step. Drop it in a directory the agent watches and it applies that procedure when the task matches.

The difference is between hoping the agent does the right thing and telling it exactly what right means for your codebase.

The five that changed my workflow

feature-plan - before writing any code, produce a plan: what changes, which files, what could break. Skipping this is how you get 400 lines of code nobody wanted.

test-write - write the test first, then make it pass. The skill encodes the pattern and makes the agent follow it instead of retrofitting tests onto code it already wrote.

code-review - a review pass with a specific checklist. Catches the things you skim past at 1am: unhandled null, a missing await, an N+1 query that won't matter until you have 10,000 users.

bug-fix - reproduce, isolate, fix, verify. In that order. The most common waste in agent-assisted debugging is changing code before you can reproduce the bug, then changing it again.

refactor - safe refactoring with verification at each step, so cleanup doesn't become an outage.

Wiring it up

Drop the folder in. Point your agent at the skills directory. For Claude Code the config is a JSON file; for Cursor there's a .cursorrules file. Both are in the pack.

The real payoff isn't the individual procedures. It's that they're consistent, so your agent stops improvising a different approach every session.

What this doesn't replace

Type hints, tests, and code review. Skills make your agent more consistent, not more competent. A skill that says "write good code" is useless; a skill that says "before editing any file, read CLAUDE.md and follow the existing naming convention" changes behavior.

Be specific or be ignored.

The one that gets skipped

Nobody wires up their MCP servers and skills on day one. They install the editor, write a prompt, feel productive, and never revisit it.

The setup takes twenty minutes. Not doing it costs you the same three weeks of drift every three weeks, forever.

I packaged these five skills plus a Cursor config, project rules, and starter templates into the MCP Vibe Coder Pack:

MCP Vibe Coder Pack - https://baradigitaltools.gumroad.com/l/mcp-vibe-coder-pack

There's also a bundle that adds 20 freelance prompts and a remote jobs dataset, if you're trying to turn the shipping into income:

Vibe Coder Power Pack - https://baradigitaltools.gumroad.com/l/vibe-coder-power-pack

Your Software Engineer Resume Gets Rejected Before Anyone Reads It

The Android AI Architect — Mon, 05 Oct 2026 09:22:10 +0000

Most software engineers don't have a resume problem. They have a parsing problem.

Your resume goes through an Applicant Tracking System that converts it to plain text, scores it against the job description, and ranks it. Only the top 10-15% reach a human. If you're applying hard and hearing nothing, this is why.

What the filter actually sees

Three layers decide your fate:

Machine layer - ATS parsing. Two-column layouts, tables, icons, and graphics often break parsers. Your contact info lands in the wrong field. Your dates vanish.
Semantic layer - AI keyword scoring. The parser compares your text to the job description. Vague bullets score nothing.
Human layer - the 7-second recruiter scan. Skimmed on a phone.

Most resumes fail layer one and never reach layer three.

The rewrite that changes outcomes

Before:

Worked on React components
Helped backend team
Responsible for bug fixes

After:

Developed 90+ reusable React components, cutting page load 40%
Architected REST APIs, reducing p95 response time by 150ms
Increased test coverage to 85%, reducing production incidents 30%

Same work. The second version has numbers, ownership, and outcome. That's what a parser can score and a recruiter can scan.

The structure that survives parsing

Single column. No tables, no text boxes, no headers or footers.
Standard section names: Experience, Education, Skills.
PDF, not a decorative template.
Bullets start with a verb: Developed, Architected, Reduced, Increased.
Every bullet answers: what did you do, how, and what changed?

Keywords without keyword stuffing

Read the job description. If it says Kubernetes and your resume doesn't, you're invisible to the semantic layer, regardless of skill.

Mirror their exact phrasing where it's true. "Built REST APIs with Node.js and PostgreSQL" beats "Backend developer" even if they mean the same thing.

The mistake that costs the most

Most engineers rewrite for aesthetics. They pick a nicer template, add a skills bar, compress to one page, and get the same rejections.

Optimization is structural, not cosmetic. Fix the parse, then fix the bullets, then worry about how it looks.

Fix the right thing first

If your applications aren't converting, the bottleneck might not be your resume at all. Your resume could be fine while your targeting is wrong, or your application volume is too low, or your interview technique is costing you offers you already earned.

A diagnostic tells you which one before you spend a month fixing the wrong thing.

I bundled the ATS template, the before/after case study, the cover letter pack, and that diagnostic into one system:

Job Hunt Complete Bundle - https://baradigitaltools.gumroad.com/l/job-hunt-complete-bundle

The ATS Resume Template and AI Optimization System is also available on its own.

DEV Community: The Android AI Architect