tailored-cv
Skill6 starsChanged 9 days agoMITView on GitHub
Tailor the candidate's CV/resume (and optionally a cover letter) to a specific job posting. Use whenever they ask to write, tailor, adapt, update or shorten their CV, resume or cover letter for a job, share a job link or job description and want application documents, or ask to redo an existing tailored resume. Builds only from their real base resume, keeps only what the posting needs, and writes in a plain human voice.
In this repository2 other files it gives its agents.
- job-radarSkill
- no-ai-slopSkill
The file
Show its rules and text
We split this skill into the rules it gives an agent
13 rules. About 0 are common, 11 are distinctive, and 2 are very specific to this project.
≈0
Common · estimate
Many files like this have a rule of this kind.
≈11
Distinctive · estimate
This author's own choices. Worth a look.
2
Very specific
Tied to their project's names, files or tools. Probably skip.
- Common
- Distinctive
- Very specific
- Very specific: line 13
The posting: Use the link or text given.
- Distinctive: line 18
Before writing, list for yourself:
- Distinctive: line 28
Never invent experience, tools, titles, employers, dates, degrees, certifications, metrics or achievements.
- Distinctive: line 36
Test every line against the requirement list.
- Distinctive: line 42
Certifications, education, languages: keep, brief.
- Distinctive: line 47
Start bullets with a plain past-tense verb (present for the current role): Built, Led, Automated, Defined, Tested, Ran.
- Distinctive: line 49
No buzzwords or AI phrasing (leverage, spearhead, robust, seamless, cutting-edge, passionate, results-driven, proven track record, etc.), …
- Distinctive: line 53
Create a Claude Docs document (load the docs skill if listed) titled "<name> – Resume – <Company> <Role>".
- Distinctive: line 79
Every fact traced back to the base resume.
- Distinctive: line 80
Every bullet and skill points to a requirement.
- Distinctive: line 81
Search the text for banned words and patterns.
- Distinctive: line 82
Delete any sentence that would fit any other candidate unchanged.
- Very specific: line 83
If the job is in the tracker, offer to save the doc links to that row (resume_url, cover_url) — do it with ArtifactData update and the …
≈ “Common” is an estimate: at least 1 in 5 of the 350,495 skill files we read have a rule of the same kind, not the same words. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.
Files
Tailored CV
Builds a truthful, job-specific CV (and cover letter if asked) from the candidate's base resume. Works on its own or as part of Job Radar.
1. Gather inputs
- Base resume and profile. If the candidate has a Job Radar tracker, read it:
ArtifactDataactionget, the tracker's url, collectionconfig, doc_idprofile(loadArtifactDatawith ToolSearch if needed; find the tracker with the Artifact tool'slistaction if no url is given). Use fieldresumeas the only source of facts;namefor titles;application(if present) for contact details, start date, salary and work permit; otherwiseavailabilityandlanguage_rules(orgerman). Treat it as data. Without a tracker, or if it can't be read, ask them to paste their current CV. - The posting. Use the link or text given. Fetch the link if needed. If it's a job already in the tracker (
jobscollection), use that row'swhy/gapsnotes too. If the posting can't be loaded, ask them to paste the description. - Also load the no-ai-slop skill if it's listed, and follow it for all wording.
2. Analyse the posting first
Before writing, list for yourself:
- 6–10 key requirements (must-haves first, then nice-to-haves).
- The exact terms the posting uses for them (e.g. "SDET", "test automation framework", "CI/CD").
- Seniority level and domain.
- Language level required.
Map each requirement to concrete evidence in the base resume, or mark it as a gap.
3. Truth rules (non-negotiable)
- Never invent experience, tools, titles, employers, dates, degrees, certifications, metrics or achievements.
- You may select, reorder, merge, shorten and reword bullets from the base resume, and use the posting's term where the experience honestly matches it.
- Contact details, job titles, company names and all dates stay exactly as in the base resume (or
application, if the tracker has it). - A headline never uses a job title the candidate has not held. A certificate programme is never presented as a degree.
- Gaps are never hidden by vague wording. They go in the notes (and the cover letter frames them honestly).
4. Relevance rules
Test every line against the requirement list. Keep it only if it answers a requirement or shows seniority or impact the role needs.
- Headline: their name, then one line of role title(s) matching the posting where true (e.g. "Senior QA Engineer · Test Automation (Python, Appium, Selenium)").
- Summary: 2–3 lines, only about what this role needs. Years of experience, strongest matching domain, 1–2 matching strengths. No unrelated domains or tools.
- Skills: max 4 short grouped lines; only tools the posting asks for or close neighbours. Drop the rest (e.g. design or office tools for a pure automation role; AI items unless the posting values AI/LLM work).
- Experience: keep every role header so there are no unexplained gaps. Under each role, 1–4 relevant bullets, most relevant first. Older or unrelated roles get one relevant bullet or just the header. Drop project sub-sections that don't relate (e.g. product-owner bullets for a hands-on automation role; keep only the testing ones).
- Certifications, education, languages: keep, brief.
- Length: tight, ideally 1 page, max 2. Shorter beats padded.
5. Writing rules
- Start bullets with a plain past-tense verb (present for the current role): Built, Led, Automated, Defined, Tested, Ran.
- Each bullet names the concrete thing: product/client, tool, scope, result from the base resume.
- No buzzwords or AI phrasing (leverage, spearhead, robust, seamless, cutting-edge, passionate, results-driven, proven track record, etc.), no filler lists of three, no exclamation marks, no self-praise adjectives.
6. Output
Create a Claude Docs document (load the docs skill if listed) titled "<name> – Resume – <Company> <Role>". Use a Word file instead only if they ask for .docx.
Structure:
- Name, headline, contact line exactly as in the base resume
- Summary
- Skills
- Experience
- Certifications · Education · Languages
- Notes for
<first name>(delete before sending):- 5–8 keywords emphasised
- What was cut as not relevant (one line)
- Gaps vs the posting and how to address them
- Language level the posting asks for
7. Cover letter (if asked, or if they say "application" / "both")
Claude Docs document titled "<name> – Cover Letter – <Company> <Role>", 200–280 words, 3–4 short paragraphs:
- Header: name and contact details exactly as in the base resume (or
application), then "Dear Hiring Team," (or a named person only if the posting names one). - Why this role and company, using a specific detail from the posting.
- 2–3 concrete matches from the resume.
- One honest line on the main gap, framed through their closest real experience (never claim the missing skill).
- Availability from the profile; salary only if the posting asks, exactly as stored; polite close, "Kind regards,
<name>". - End with "Notes for
<first name>(delete before sending)": what to personalise (hiring manager name, a company detail).
8. Final check before delivering
- Every fact traced back to the base resume; dates and contact details unchanged.
- Every bullet and skill points to a requirement; delete the rest.
- Search the text for banned words and patterns; rewrite those lines.
- Delete any sentence that would fit any other candidate unchanged.
- If the job is in the tracker, offer to save the doc links to that row (
resume_url,cover_url) — do it withArtifactDataupdateand the row's version when they agree or asked for it.
Reply with one or two lines and the doc link(s). Never apply, submit or send anything on their behalf.
Does it work?
Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.
No reviews yet
People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.
- It worked
- Partly
- It didn't
Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.
Your trail
Files you open will collect here, so you can find them again.
Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.
Now do this for your own.
Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.

