# Resume Builder

> >

- **Type:** Skill
- **Install:** `agentstack add skill-sriram789-agentic-skills-resume-builder`
- **Verified:** Yes — security-reviewed for prompt injection and unsafe behavior
- **Seller:** [sriram789](https://agentstack.voostack.com/s/sriram789)
- **Installs:** 0
- **Category:** [Agent Skills](https://agentstack.voostack.com/c/agent-skills)
- **Latest version:** 0.1.0
- **License:** Apache-2.0
- **Upstream author:** [sriram789](https://github.com/sriram789)
- **Source:** https://github.com/sriram789/agentic-skills/tree/staging/skills/resume/resume-builder

## Install

```sh
agentstack add skill-sriram789-agentic-skills-resume-builder
```

Requires the [AgentStack CLI](https://agentstack.voostack.com/docs/cli). Works with Claude Code, Cursor, and any MCP-compatible agent.

## About

# ATS-Optimized Resume Builder

Your goal is to produce a resume that does two things well: (1) pass ATS (Applicant Tracking System) filters at big tech companies, and (2) compel a human recruiter or hiring manager to pick up the phone. These goals are complementary — a keyword-rich, achievement-dense resume wins on both fronts.

## ATS-first principles (understand these before writing a single word)

ATS systems at Google, Meta, Amazon, and similar companies parse resumes before a human ever sees them. A resume that looks beautiful but fails ATS is invisible. The rules:

- **Single-column layout only** — multi-column, tables used for layout, and text boxes confuse parsers
- **Standard section names** — use exactly: Summary, Experience, Skills, Projects, Education, Certifications
- **No headers/footers for content** — ATS often ignores header/footer regions entirely
- **Standard fonts** — Arial or Calibri, 10–12pt body, never embed fonts as images
- **No graphics, icons, or images** — including profile photos, skill bars, rating dots
- **Keywords must appear in context** — "Java" in a bullet about what you built beats "Java" in a skills list
- **Bullet points, not paragraphs** — ATS scores keyword density; short bullets score higher than prose
- **Dates in consistent format** — "Jan 2023 – Present" or "2021 – 2023", never mixed

## Step 1: Read the input files

The user points you to a folder. Read every file in it — performance reviews, self-assessments, award nominations, project write-ups, previous resumes, anything. Extract:
- Job titles, companies, employment dates
- Responsibilities, achievements, projects
- Quantifiable results (numbers, percentages, scale, time savings)
- Technologies, tools, frameworks, methodologies
- Certifications, courses, awards

## Step 2: Gather targeting information

Ask these questions — they directly determine keyword strategy and positioning:

**Batch 1 (required):**
- What specific role are you targeting? (e.g., "Lead Software Engineer", "Staff Backend Engineer", "Principal Engineer")
- Which companies are highest priority? (This shapes keyword emphasis — Google loves "scalability" and "distributed systems"; Amazon loves "ownership" and "impact at scale"; Meta loves "growth" and "infrastructure")
- Do you have a job description you'd like to optimise against? If yes, paste it or share the file — this is the single highest-leverage input for ATS optimisation.
- What's your preferred output format — `.docx`, `.pdf`, or both?

**Batch 2 (fill gaps from files):**
- What's the most impressive project that might not be obvious from documents?
- Any open source contributions, public talks, or published writing?
- Preferred resume length — 1 page or 2 pages? (For 8+ years of experience targeting senior roles, 2 pages is standard and expected at big tech.)

## Step 3: Keyword strategy

Before writing, build a keyword list from two sources:

1. **From the job description** (if provided): extract every technical term, tool, methodology, and soft-skill phrase that appears. These become required keywords.
2. **From big tech job patterns for Lead/Staff roles**: regardless of JD, these clusters consistently appear in ATS filters:
   - *Backend*: distributed systems, microservices, system design, high availability, fault tolerance, scalability, low latency
   - *AI/ML*: LLM, AI agents, machine learning, model deployment, inference, embeddings, RAG, prompt engineering
   - *Infrastructure*: Kafka, Kubernetes, Docker, AWS/GCP/Azure, CI/CD, observability, Prometheus, Grafana
   - *Leadership*: technical leadership, cross-functional, mentorship, architecture, roadmap, stakeholder

Map these keywords to specific experiences in the user's background. Every keyword should appear in a bullet, not just in the Skills section.

## Step 4: Rewrite bullets for maximum impact

This is where most resumes fail. Transform every bullet using this formula:

**Formula:** `[Strong verb] + [what you built/did] + [scale/context] + [quantified result]`

Weak: "Worked on backend services for the AI team"
Strong: "Architected event-driven Kafka pipeline processing 2M+ daily AI agent interactions, reducing end-to-end latency by 40%"

Weak: "Improved debugging process"
Strong: "Built AI Thread Analyser tool reducing agent debugging time from 45 minutes to <5 minutes; adopted by 3+ teams, cutting L2 escalations by 60%"

**Strong verb bank for senior engineering roles:**
Architected · Designed · Led · Owned · Drove · Scaled · Optimised · Reduced · Eliminated · Launched · Delivered · Migrated · Automated · Standardised · Mentored · Enabled

**Rules for bullets:**
- Start with a past-tense strong verb (even for current role — ATS prefers consistency)
- Include at least one number per bullet (users, requests, %, time, $, team size)
- Keep to one line where possible; two lines maximum
- If you don't have exact numbers, use informed estimates with context: "~2M daily", "50% reduction", "10+ engineers"
- 4–6 bullets per role is the sweet spot for senior roles

## Step 5: Build the resume structure

Use this exact section order and naming:

```
[FULL NAME]
[City, Country] · [Email] · [Phone] · [LinkedIn URL] · [GitHub URL]

SUMMARY
[3–4 lines maximum]

EXPERIENCE
[Role Title] · [Company] · [Start – End]
• bullet
• bullet

SKILLS
[Grouped by category, comma-separated, no ratings or bars]

PROJECTS  (include if projects are notable and distinct from work experience)
[Project Name] · [Tech stack] · [Date/duration]
• bullet

EDUCATION
[Degree] · [Institution] · [Year]

CERTIFICATIONS  (if any)
```

### Summary section

The summary is the only prose section. It should be 3–4 lines that:
1. State the role you're targeting and your years of experience
2. Name 2–3 specific technical domains you're strongest in
3. Include one standout achievement with a number
4. End with what you're looking for

It functions as an executive summary for the recruiter *and* as a dense keyword block for ATS. Don't make it generic — make it specific enough that it could only describe this person.

### Skills section

Group skills into 3–4 categories. Keep it flat — no ratings, no bars, no icons:

```
Languages: Java, Python, JavaScript, TypeScript, SQL
Backend & Distributed Systems: Spring Boot, Apache Kafka, Elasticsearch, Redis, Microservices, System Design
AI / LLM: LLM Integration, AI Agents, Prompt Engineering, MCP (Model Context Protocol), RAG, Azure OpenAI
Cloud & DevOps: AWS, Docker, Kubernetes, ArgoCD, Prometheus, Grafana, CI/CD
```

## Step 6: Generate the document

### For .docx output
Use `docx` npm package. Install with `npm install docx` in the output directory if needed.

**Critical ATS formatting rules for docx:**
- Single section, no columns
- No tables used for layout (tables for data only, if needed at all)
- No text boxes
- No headers/footers
- Font: Arial or Calibri, 10.5pt body, 11pt for role titles, 13–14pt for name
- Margins: 0.75 inches all sides (gives space while staying ATS-safe)
- Bullet character: use `LevelFormat.BULLET` with `•` — never unicode manually
- Name: bold, 20–22pt, left-aligned
- Section headers: bold, 11pt, ALL CAPS, with a bottom border line for visual separation
- Line spacing: 1.0–1.15, with modest spacing between sections
- No color except black/dark gray (some ATS strip color; keep it safe)

### For .pdf output
Convert the .docx to PDF using LibreOffice:
```bash
python /path/to/scripts/office/soffice.py --headless --convert-to pdf resume.docx
```

### File naming
Name output files clearly:
- `[FirstName]_[LastName]_Resume_[TargetRole].docx`
- `[FirstName]_[LastName]_Resume_[TargetRole].pdf`

## Step 7: ATS self-check before delivering

Before handing the resume to the user, mentally verify:
- [ ] Every section uses the standard name (Experience, not "Work History")
- [ ] All dates are consistent format
- [ ] At least 8–10 of the target keywords appear in bullet context (not just Skills)
- [ ] No bullet starts with "I" or "We" or "Responsible for"
- [ ] Every role has at least one quantified achievement
- [ ] No graphics, tables-as-layout, or multi-column formatting
- [ ] Skills section is complete and grouped

## Tone guidance

Be assertive. The user wants maximum reach — don't hedge or soften. If a bullet is weak, rewrite it strong and explain why. If a section is missing something important for big tech, say so and add it. The resume should read like the work of someone who is clearly ready for Lead/Staff — decisive, ownership-oriented, impact-focused.

## Source & license

This open-source skill is cataloged on AgentStack and links to its original source — we do not rehost the code.

- **Author:** [sriram789](https://github.com/sriram789)
- **Source:** [sriram789/agentic-skills](https://github.com/sriram789/agentic-skills)
- **License:** Apache-2.0

Install and usage instructions live in the source repository linked above.

## Pricing

- **Free** — Free

## Security capabilities

Automated source analysis of v0.1.0 — what this tool can access:

- **Network access:** no
- **Filesystem access:** no
- **Shell / process execution:** no
- **Environment & secrets:** no
- **Dynamic code execution:** no

*"Yes" means the capability is present in the source — more access means more to trust, not that it is unsafe.*


## Versions

- **0.1.0** — security scan: passed — Imported from the upstream source.

## Links

- Listing page: https://agentstack.voostack.com/l/skill-sriram789-agentic-skills-resume-builder
- Seller: https://agentstack.voostack.com/s/sriram789
- Browse the marketplace: https://agentstack.voostack.com/browse

---
Listed on AgentStack — the marketplace for AI agent skills and MCP servers. Every listing is security-reviewed. Creators keep 70%.