Technical job search
Turn a concrete job-search input such as a job description, resume, offer, or interview context into a structured hiring artifact that preserves the user's evidence, reduces generic prose, and produces decision-ready application material.
When to invoke
- "Analyze this job description for must-haves and red flags."
- "Tailor my CV to this software engineer role."
- "Write a cover letter for this job application."
- "Evaluate this job offer and compensation package."
- "Draft a follow-up email after my interview."
Inputs
Use $ARGUMENTS to determine the requested job-search action when it is provided. If $ARGUMENTS is empty, infer the action from the user's pasted material and ask only for missing essentials: job description for tailoring, resume for resume edits, offer details for offer evaluation, or interview context for follow-up email.
Job description analysis
Separate explicit requirements from recruiter filler. Treat repeated items and words such as "required", "must", "minimum", and "need" as stronger signals than isolated "preferred" or "bonus" phrases.
| Extract |
How to classify |
Output rule |
| Must-haves |
Required or repeated technical skills, years of experience, domain knowledge, clearance, location, work authorization, or language requirements. |
List as bullets and mark any hard blockers. |
| Nice-to-haves |
Preferred, bonus, familiar-with, or mentioned once without being required. |
Keep separate so the candidate does not self-reject. |
| Role problem |
The business or engineering problem the hire is meant to solve. |
Summarize in 2-3 sentences, using evidence from the JD. |
| Seniority signals |
Verbs such as build, design, lead, define, mentor, own, influence, or set strategy. |
Map to likely level and rewrite resume bullets to match that level. |
| Red flags |
Undefined ownership, impossible stack breadth, vague team or product scope, urgency without clarity. |
Surface as questions for the recruiter or hiring manager. |
Common red flags to call out:
- "Wear many hats" with no scope boundaries: risk of undefined ownership.
- 10+ must-have technologies for one role: unrealistic bar or poor planning.
- No team size, tech stack, roadmap, or shipped product: possible disorganization.
- "Fast-paced" plus after-hours language: probe on on-call and work-life expectations.
Resume and CV tailoring
Use this section for CV/resume work. Do not keyword-stuff; match terms only where the candidate has real evidence.
Rewrite for the hiring manager first and ATS second. Match real evidence from the user's background to the role; never invent experience.
| Rule |
Apply it like this |
Avoid |
| Match language exactly |
If the JD says distributed systems, use distributed systems, not large-scale systems. |
Keyword stuffing that repeats terms without evidence. |
| Lead with impact |
Start bullets with result and scale: "Reduced P99 latency by 40%". |
"Worked on performance improvements". |
| Quantify scope |
Include users, QPS, revenue, cost saved, team size, incident volume, or latency. |
Unmeasured claims such as "improved reliability". |
| Cut unrelated detail |
Keep a two-page tailored CV over a four-page generic CV. |
Keeping every old technology because it is impressive. |
| Mirror seniority |
Entry: built; senior: designed; staff/principal: drove, defined, influenced. |
Overstating ownership beyond the evidence. |
Use this bullet formula when possible: Verb + technical object + method/constraint + measurable result. Example: Reduced checkout P99 latency by 40% by moving payment retries to an idempotent queue-backed workflow.
Cover letters and follow-up emails
Keep cover letters under 300 words in three short paragraphs, with no preamble such as "I am writing to apply for" and no prose restatement of the CV.
| Artifact |
Required content |
Hard limit |
| Cover letter paragraph 1 |
Why this company: product, technical challenge, published work, or problem space. |
Be specific; never say only "I admire your mission". |
| Cover letter paragraph 2 |
Why this candidate: one or two concrete background matches to the role. |
Do not summarize the full career. |
| Cover letter paragraph 3 |
Why now: one sentence on motivation at this career point. |
Keep the whole letter under one page. |
| Follow-up email |
Thank them, reference one specific conversation detail, reaffirm genuine interest. |
Send within 24 hours; no multi-paragraph recap. |
Avoid "I am passionate about", "leader in the industry", "innovative company", and generic praise that could apply to any employer.
Offer evaluation
Compare total compensation and role risk, not just base salary. Use market sources only as reference points and explain uncertainty when data is stale or sparse.
| Dimension |
Questions to answer |
Evidence to request |
| Compensation |
Base vs market, target bonus vs historical payout, equity annualized value, refreshers. |
Offer letter, vesting schedule, strike price or valuation context, bonus plan. |
| Equity |
Four-year vesting with one-year cliff is common; early-stage equity has dilution and liquidity risk. |
Share count, fully diluted shares, latest valuation, exercise window. |
| Role clarity |
What does "own" mean, what is already decided, who sets priorities? |
Hiring manager answers, org chart, first-90-days plan. |
| Growth |
What does next level require and how long do people typically take? |
Leveling rubric, promotion examples, manager expectations. |
| Company health |
Runway, revenue, growth, customer concentration, funding stage. |
Written answers where possible, public filings or credible market data. |
| Engineering culture |
PR review, incident postmortems, on-call load, remote/hybrid reality. |
Team interview notes, written policy, sample on-call rotation. |
Offer red flags: pressure to decide in under 48 hours, equity with no liquidity path after 10+ private years, greenfield claims hiding unmaintained code, or refusal to put material terms in writing. Get everything in writing before accepting.
Gotchas
- Do not invent credentials or achievements: tailor phrasing only from the user's real resume, notes, or interview details.
- Do not optimize only for ATS: keyword matching matters, but the hiring manager must still see impact and evidence.
- Do not treat nice-to-haves as blockers: many candidates self-select out because optional phrases are mixed with requirements.
- Do not give legal or financial advice: explain compensation mechanics and risk questions, and recommend a qualified advisor for contract, tax, visa, or equity exercise decisions.
Output template
## Job-search result — <task>
**Status:** ready | needs input | blocked
**Target role/company:** <role and company, or "not provided">
**Source material reviewed:** <JD/resume/offer/interview notes>
### Key findings
- <must-have, match, risk, or decision point>
- <must-have, match, risk, or decision point>
### Draft artifact
<tailored bullets, cover letter, follow-up email, offer comparison, or JD analysis>
### Open questions
- <question to ask recruiter, hiring manager, or user>
Quality gate
1---2name: technical-job-search-33description: Help software engineers perform active job-search tasks: analyze job descriptions, tailor resumes, write concise cover letters, evaluate offers, and draft interview follow-ups. Use when the user asks to parse a JD, match a CV to a role, write application material, compare compensation, assess offer risk, or prepare post-interview outreach.4license: MIT5---67<!-- Generated from harness/github-copilot/plugins/career-productivity/skills/technical-job-search/SKILL.md by harness/claude-code/scripts/convert_from_copilot.py. Edit the source, not this file. -->89# Technical job search1011Turn a concrete job-search input such as a job description, resume, offer, or interview context into a structured hiring artifact that preserves the user's evidence, reduces generic prose, and produces decision-ready application material.1213## When to invoke1415- "Analyze this job description for must-haves and red flags."16- "Tailor my CV to this software engineer role."17- "Write a cover letter for this job application."18- "Evaluate this job offer and compensation package."19- "Draft a follow-up email after my interview."2021## Inputs2223Use `$ARGUMENTS` to determine the requested job-search action when it is provided. If `$ARGUMENTS` is empty, infer the action from the user's pasted material and ask only for missing essentials: job description for tailoring, resume for resume edits, offer details for offer evaluation, or interview context for follow-up email.2425## Job description analysis2627Separate explicit requirements from recruiter filler. Treat repeated items and words such as "required", "must", "minimum", and "need" as stronger signals than isolated "preferred" or "bonus" phrases.2829| Extract | How to classify | Output rule |30| --- | --- | --- |31| Must-haves | Required or repeated technical skills, years of experience, domain knowledge, clearance, location, work authorization, or language requirements. | List as bullets and mark any hard blockers. |32| Nice-to-haves | Preferred, bonus, familiar-with, or mentioned once without being required. | Keep separate so the candidate does not self-reject. |33| Role problem | The business or engineering problem the hire is meant to solve. | Summarize in 2-3 sentences, using evidence from the JD. |34| Seniority signals | Verbs such as build, design, lead, define, mentor, own, influence, or set strategy. | Map to likely level and rewrite resume bullets to match that level. |35| Red flags | Undefined ownership, impossible stack breadth, vague team or product scope, urgency without clarity. | Surface as questions for the recruiter or hiring manager. |3637Common red flags to call out:3839- "Wear many hats" with no scope boundaries: risk of undefined ownership.40- 10+ must-have technologies for one role: unrealistic bar or poor planning.41- No team size, tech stack, roadmap, or shipped product: possible disorganization.42- "Fast-paced" plus after-hours language: probe on on-call and work-life expectations.4344## Resume and CV tailoring4546Use this section for `CV/resume` work. Do not `keyword-stuff`; match terms only where the candidate has real evidence.474849Rewrite for the hiring manager first and ATS second. Match real evidence from the user's background to the role; never invent experience.5051| Rule | Apply it like this | Avoid |52| --- | --- | --- |53| Match language exactly | If the JD says `distributed systems`, use `distributed systems`, not `large-scale systems`. | Keyword stuffing that repeats terms without evidence. |54| Lead with impact | Start bullets with result and scale: "Reduced P99 latency by 40%". | "Worked on performance improvements". |55| Quantify scope | Include users, QPS, revenue, cost saved, team size, incident volume, or latency. | Unmeasured claims such as "improved reliability". |56| Cut unrelated detail | Keep a two-page tailored CV over a four-page generic CV. | Keeping every old technology because it is impressive. |57| Mirror seniority | Entry: built; senior: designed; staff/principal: drove, defined, influenced. | Overstating ownership beyond the evidence. |5859Use this bullet formula when possible: `Verb + technical object + method/constraint + measurable result`. Example: `Reduced checkout P99 latency by 40% by moving payment retries to an idempotent queue-backed workflow`.6061## Cover letters and follow-up emails6263Keep cover letters under 300 words in three short paragraphs, with no preamble such as "I am writing to apply for" and no prose restatement of the CV.6465| Artifact | Required content | Hard limit |66| --- | --- | --- |67| Cover letter paragraph 1 | Why this company: product, technical challenge, published work, or problem space. | Be specific; never say only "I admire your mission". |68| Cover letter paragraph 2 | Why this candidate: one or two concrete background matches to the role. | Do not summarize the full career. |69| Cover letter paragraph 3 | Why now: one sentence on motivation at this career point. | Keep the whole letter under one page. |70| Follow-up email | Thank them, reference one specific conversation detail, reaffirm genuine interest. | Send within 24 hours; no multi-paragraph recap. |7172Avoid "I am passionate about", "leader in the industry", "innovative company", and generic praise that could apply to any employer.7374## Offer evaluation7576Compare total compensation and role risk, not just base salary. Use market sources only as reference points and explain uncertainty when data is stale or sparse.7778| Dimension | Questions to answer | Evidence to request |79| --- | --- | --- |80| Compensation | Base vs market, target bonus vs historical payout, equity annualized value, refreshers. | Offer letter, vesting schedule, strike price or valuation context, bonus plan. |81| Equity | Four-year vesting with one-year cliff is common; early-stage equity has dilution and liquidity risk. | Share count, fully diluted shares, latest valuation, exercise window. |82| Role clarity | What does "own" mean, what is already decided, who sets priorities? | Hiring manager answers, org chart, first-90-days plan. |83| Growth | What does next level require and how long do people typically take? | Leveling rubric, promotion examples, manager expectations. |84| Company health | Runway, revenue, growth, customer concentration, funding stage. | Written answers where possible, public filings or credible market data. |85| Engineering culture | PR review, incident postmortems, on-call load, remote/hybrid reality. | Team interview notes, written policy, sample on-call rotation. |8687Offer red flags: pressure to decide in under 48 hours, equity with no liquidity path after 10+ private years, greenfield claims hiding unmaintained code, or refusal to put material terms in writing. Get everything in writing before accepting.8889## Gotchas9091- **Do not invent credentials or achievements**: tailor phrasing only from the user's real resume, notes, or interview details.92- **Do not optimize only for ATS**: keyword matching matters, but the hiring manager must still see impact and evidence.93- **Do not treat nice-to-haves as blockers**: many candidates self-select out because optional phrases are mixed with requirements.94- **Do not give legal or financial advice**: explain compensation mechanics and risk questions, and recommend a qualified advisor for contract, tax, visa, or equity exercise decisions.9596## Output template9798```markdown99## Job-search result — <task>100101**Status:** ready | needs input | blocked102**Target role/company:** <role and company, or "not provided">103**Source material reviewed:** <JD/resume/offer/interview notes>104105### Key findings106- <must-have, match, risk, or decision point>107- <must-have, match, risk, or decision point>108109### Draft artifact110<tailored bullets, cover letter, follow-up email, offer comparison, or JD analysis>111112### Open questions113- <question to ask recruiter, hiring manager, or user>114```115116## Quality gate117118- [ ] The response stays tied to an active job-search action, not broad career coaching.119- [ ] Must-haves and nice-to-haves are separated when analyzing a JD.120- [ ] Resume bullets preserve the user's actual experience and add no invented facts.121- [ ] Cover letters stay under 300 words and avoid generic company praise.122- [ ] Offer evaluation includes compensation, role clarity, company health, and written-term risks.123- [ ] Follow-up emails are concise and reference a real interview detail.124- [ ] The output follows `## Output template` exactly.
paulasilvatech