splash0047/resume-generator / .agents/skills/jd-intelligence-analyzer
JD Intelligence Analyzer
VerifiedTransform a raw job description into a weighted keyword requirement graph and infer the writing style profile from JD content (not company name). Use this before any resume generation task.
JD Intelligence Analyzer
When to Use This Skill
Use this skill as Phase 0 — the very first step — whenever the user provides a job description. This skill must run before:
- Resume tailoring
- Bullet generation
- Evidence database population
- Humanizer style selection
Trigger phrases: "I have a JD", "here's the job posting", "I'm applying to", "optimize for this role", "tailor my resume"
Core Capabilities
- Build a weighted keyword requirement graph from any job description
- Classify keywords into importance tiers (Critical / Major / Minor)
- Infer the correct writing style profile from JD content — never from company name
- Detect company type and audience tier from verbs, responsibilities, and domain terms
- Output a structured briefing that all downstream pipeline stages consume
Step 1: Keyword Importance Graph
Parse the full JD text and extract every technical term, skill, methodology, and domain word. Then assign each an importance tier based on frequency and placement.
Importance Tiers
| Tier | Criteria | Resume Impact |
|---|---|---|
| Critical | ≥5 mentions OR appears in "Required" / "Must have" / "You have" | MUST appear in resume. Missing = likely rejection. |
| Major | 3–4 mentions OR appears in "Preferred" / "Nice to have" / "Bonus" | SHOULD appear. Absence is a significant gap. |
| Minor | 1–2 mentions, not emphasized | Nice to include. Absence is acceptable. |
Output Format
## Keyword Importance Graph
### Critical Keywords (must appear in resume)
| Keyword | Mentions | JD Section |
|---|---|---|
| Python | 12 | Required, Responsibilities (×8), Preferred |
| FastAPI | 7 | Required, Tech Stack, Responsibilities |
| Docker | 5 | Required, Tech Stack |
### Major Keywords (should appear)
| Keyword | Mentions | JD Section |
|---|---|---|
| Kubernetes | 4 | Preferred, Tech Stack |
| PostgreSQL | 3 | Preferred |
### Minor Keywords (nice to have)
| Keyword | Mentions | JD Section |
|---|---|---|
| Redis | 1 | Bonus |
| GraphQL | 1 | Bonus |
### Weighted Match Score Formula
Overall Match = (Critical Present% × 0.60) + (Major Present% × 0.30) + (Minor Present% × 0.10)
Target: ≥85% overall
Step 2: Writing Style Inference (2D Matrix)
Infer the writing style from JD content only. Never use the company name as the primary signal — the same company may post very different roles requiring very different styles.
Dimension 1: Audience Tier
Detect the primary reader of the resume from the responsibilities and interview framing in the JD.
| Audience | JD Signals |
|---|---|
| Recruiter | High-level role descriptions, broad skill lists, "culture fit" language, salary mentioned, benefits emphasized |
| Engineering Manager | "You will design / lead / architect", code review mentions, system design references, team collaboration language |
| Researcher | "Co-author", "publish", "benchmark", "experiment", "literature review", conference mentions |
Dimension 2: Company Type
Detect from domain vocabulary, product references, and action verbs.
| Company Type | JD Signals |
|---|---|
| Startup / Open Source | "Ship", "own it", "iterate", "move fast", "small team", "wear many hats", "open-source contribution", "PRs", "community" |
| Enterprise / Big Tech | "Scale", "SLA", "compliance", "stakeholder", "roadmap", "10M+ users", "distributed systems", "reliability", "on-call" |
| Research | "Training", "model", "fine-tuning", "inference", "paper", "benchmark", "experiment", "dataset", "GPU", "CUDA" |
| DevTools / API-first | "Developer experience", "SDK", "API design", "DX", "documentation", "adoption", "integration", "CLI" |
The 2D Writing Style Matrix
| Startup / Open Source | Enterprise / Big Tech | Research | DevTools / API-first | |
|---|---|---|---|---|
| Recruiter | Concise, action verbs, shipped products | Scale + reliability + team impact | Clear summaries of results | Adoption metrics, integrations |
| Engineering Manager | Architecture decisions, velocity, ownership | System design, fault tolerance, SLA | Experiment design, methodology | API contracts, extensibility, DX |
| Researcher | Novel approaches, contribution to field | Scalable research infra | Rigorous methodology, benchmarks | Research tools, reproducibility |
Inferred Profile Output
## Inferred Writing Style Profile
**Audience Tier:** Engineering Manager
**Company Type:** Startup / Open Source
**Combined Profile:** [Startup/OS × Eng Manager]
**Writing Guidance:**
- Lead with architecture decisions and technical reasoning
- Emphasize ownership and end-to-end delivery
- Open-source contributions are high signal — cite GitHub activity
- Avoid corporate language; write as an engineer, not a marketer
- Use verbs: built, designed, shipped, contributed, optimized, measured
Step 2.5: Role Family Classification
Classify the JD into a role family. This classification drives project selection and skill ordering from the Candidate Position Brief.
Role Family Detection
| Role Family | JD Signals | |---|---|---| | SDE / Software Engineer | "DSA", "algorithms", "data structures", "system design", "OOP", "clean code", language-specific requirements (Java, C++, Python), "coding round" | | AI/ML Engineer | "machine learning", "deep learning", "NLP", "LLMs", "model training", "inference", "RAG", "AI agents", "MLOps", "data pipeline" | | Full Stack Developer | "React", "Angular", "Vue", "Node.js", "REST API", "responsive design", "frontend + backend", "full stack" | | DevOps / Cloud Engineer | "CI/CD", "Kubernetes", "Docker", "Terraform", "cloud infrastructure", "monitoring", "SRE", "deployment" | | Data Engineer | "ETL", "Spark", "Kafka", "data warehouse", "BigQuery", "Airflow", "data pipeline", "streaming" |
Output
## JD Role Family
**Classified as:** [Role Family]
**Confidence:** [High / Medium / Low]
**Signals:** [List of JD phrases that led to classification]
**Comparison to Candidate Position Brief:**
- Candidate primary: [Role Family]
- JD role family: [Role Family]
- Alignment: [Match / Adjacent / Mismatch]
If mismatch: Consider switching to candidate's secondary role positioning.
Step 3: Red Flag Detection
Scan the JD for signals that indicate a poor role fit or unrealistic expectations.
| Red Flag | Signal | Action |
|---|---|---|
| Scope creep | "Wear many hats", "rockstar", "ninja", "10x" | Note: role may have undefined scope |
| Compensation mismatch | Extremely wide salary range (e.g. ₹8L–₹30L) | Flag for offer comparison analysis |
| Tech stack mismatch | >40% Critical keywords absent from candidate profile | Flag as stretch role; note gaps |
| Vague requirements | No specific technologies in "Required" | JD may be poorly written; proceed with caution |
Step 4: Project Selection Briefing
Based on the keyword importance graph, rank the candidate's projects by relevance:
## Project Selection Priority
For this JD, rank your projects in this order:
1. **[Project A]** — covers 4/5 Critical keywords (Python, FastAPI, Docker, PostgreSQL)
2. **[Project B]** — covers 2/5 Critical keywords (Python, PostgreSQL)
3. **[Project C]** — covers only Minor keywords (Redis)
Recommendation: Lead with Project A and Project B. Drop Project C unless space permits.
Technologies to Remove/Minimize
Based on the keyword importance graph and JD role family, output technologies from the candidate's profile that should be DE-EMPHASIZED on this resume:
## Technologies to Remove/Minimize
For this JD, the following candidate skills should NOT appear in Technical Skills
(or appear only if space permits after all relevant skills are listed):
| Technology | Reason |
|---|---|
| [Tech] | Not mentioned in JD; dilutes SDE specialization signal |
| [Tech] | Mentioned once as Minor; don't lead with it |
| [Tech] | Domain mismatch — this is an AI/ML technology for an SDE role |
**Skills section should have ≤15 technologies total.**
**Order by:** JD Critical → JD Major → JD Minor → Candidate core identity
Full Output Delivered to Downstream Stages
Every downstream skill (Bullet Generator, Humanizer, Resume Critic) receives:
## JD Intelligence Briefing
**Role:** [Title] at [Company]
**Date:** [Date]
### Keyword Importance Graph
[Critical / Major / Minor tables]
### Inferred Style Profile
Audience: [Recruiter | Eng Manager | Researcher]
Company Type: [Startup/OS | Enterprise/Big Tech | Research | DevTools]
Combined Profile: [e.g., Enterprise × Eng Manager]
### Project Priority Order
1. [Project name] — [reason]
2. [Project name] — [reason]
### Red Flags
[Any detected | None]
### Weighted Match Score (Current)
Before resume tailoring: X%
Target after tailoring: ≥85%
Rules
- Never use company name as the primary style signal. Always infer from JD content.
- Keyword counting must be exhaustive. Read the full JD text, including benefits and about-the-company sections (they sometimes contain repeated skill signals).
- Critical keywords must appear in the resume — missing even one is a significant risk.
- The briefing is consumed by every downstream stage — format it clearly so Bullet Generator, Humanizer, and Critic can reference it without re-reading the JD.
Related skills
| Skill | Stars | Updated | Works with | Status | Install |
|---|---|---|---|---|---|
| slack-gif-creatoranthropics/skillsKnowledge and utilities for creating animated GIFs optimized for Slack. Provides constraints, validation tools, and animation concepts. Use when users request animated GIFs for Slack like "make me a GIF of X doing Y for Slack." | Stars: 196k | Updated | Claude CodeCodex | Verified | Install |
| ponytailDietrichGebert/ponytailLazy senior dev mode for any coding task: the smallest change that fully solves it, and a reply a busy human understands in one read. | Stars: 193k | Updated | Claude CodeCodex | Verified | Install |
| ponytail-debtDietrichGebert/ponytailHarvest every shortcut: comment into one debt ledger, so deferrals get tracked instead of forgotten. One-shot report. | Stars: 193k | Updated | Claude CodeCodex | Verified | Install |
| ponytail-debtDietrichGebert/ponytailList every `shortcut:` comment (and older `ponytail:` ones) as a debt ledger. One-shot report, changes nothing. Use for "ponytail debt", "what did ponytail defer", "list the shortcuts", /ponytail-debt. | Stars: 193k | Updated | Claude CodeCodex | Verified | Install |