io.github.JesusCabreraReveles由 GitHub 帳號 JesusCabreraReveles 發布v0.3.3更新於 Nov 5, 2026
Honest, ATS-safe CVs and cover letters from a skill inventory you confirm yourself.
依據本機保存、有證據佐證的技能清單，產出誠實且對 ATS 友善的履歷、求職信與 LinkedIn 文案，並依職務類型分版本。
Desktop only，透過 STDIO。 STDIO 服務會啟動本機處理程序，因此需要 SourceWeft 桌面主機。
參照 儲存庫 中的啟動說明。
[CI] [License: MIT] [PyPI] [Listed on mcpservers.org]
Honest, ATS-safe CVs tailored to each role, generated by your AI agent. honest-cv is an MCP server, a CLI and a Claude Code skill. It keeps one source of truth about your experience and turns it into a PDF and DOCX for each type of role. It never claims a skill you don't have.
Many CVs that get no callbacks don't fail on design. They fail on targeting, and on details that quietly break text extraction. AI makes the opposite problem easy: a model will happily stuff a CV with keywords you can't defend in an interview. honest-cv is built around three ideas:
The server is deterministic and needs no API key. Your MCP client's model runs the conversation; honest-cv validates, builds, verifies and measures. It keeps your profile in a folder on your machine and never sends it anywhere: see Privacy for what your AI client does with it.
[A generated one-page CV for the fictional Ada Example]
honest-cv needs uv. Typst and the fonts ship inside the package,
so there is nothing else to install.
It runs on Linux (glibc) on x86_64 and ARM64, macOS on Intel and Apple
Silicon, and Windows on x64. Alpine Linux and Windows on ARM aren't supported, because the typst
package doesn't publish wheels for them.
To install the skill as well, which guides the interview and the tailoring:
Then follow Your first CV, step by step.
Add this to ~/.cursor/mcp.json, or to .cursor/mcp.json in a project:
Edit claude_desktop_config.json:
~/Library/Application Support/Claude/;%APPDATA%\Claude\.Add the same mcpServers block as in the Cursor example and restart the app. If the app can't
find uvx, use its absolute path (which uvx).
Raising a skill's level needs your own confirmation, which the server asks for through MCP elicitation. Not every client supports it yet:
This walkthrough uses Claude Code; any MCP client works the same way, only the setup differs.
1. Install uv and add honest-cv. One time only. On macOS or Linux:
~/cv-profile is the folder where your profile will live. It doesn't have to exist yet, and it should
not be inside a public repository. Install the skill too (Claude Code above).
2. Start the interview. Open Claude Code in any folder (claude) and write, in your own
language, "Interview me for my CV", or run /mcp__honest-cv__interview. The assistant keeps
your language for the whole conversation. You can answer by voice with your system's dictation
(on macOS: System Settings → Keyboard → Dictation): it reads names and numbers back to you before
saving them, since dictation mishears them. It goes like this:
You: Interview me for my CV.
Assistant: The folder
~/cv-profiledoesn't exist yet. Shall I create it? … The whole interview can take a few hours, but everything is saved as we go. A first session with your contact details, your two most recent roles and the skills checklist is enough for a first CV. Do you have a LinkedIn data export or a JSON Resume to start from?
backend-jvm + engineering-lead, or your own for any
profession. Open ~/cv-profile/checklist.md in your editor, put P, L, F or X on each line, add where
you used it after " — ", save, and tell the assistant "done". It asks for the evidence it is
missing, in groups. Then a form asks you to confirm the upgrades: only what you answer
"sí" is saved.3. Build your first CV. The assistant proposes one variant per kind of role you apply to (for example "platform" or "engineering manager"): title, summary and which bullets to show. Once you approve it:
You: Build my platform CV in English and Spanish.
The PDF and DOCX land in ~/cv-profile/out/, already checked for ATS safety. Without an AI client,
the same is uvx honest-cv build ~/cv-profile -v platform -l en.
4. Tailor it to a job post. Paste the post, or run /mcp__honest-cv__tailor_for_job:
You: Here is a job post: (paste it). How well do I fit, and what should I change?
You get the coverage of each variant, the missing keywords split into you have it but this CV
doesn't show it, familiarity only and you don't have it, plus the ones you marked in your
checklist but haven't backed with evidence yet (it asks where you used them), a tailored variant to approve, and
the build. Then it can draft a cover letter, which check_claims checks before it is built, and
log the application with the post's text and what to prepare for the interview. When you get the
interview, "help me prepare the interview" drafts likely questions with answers from your own
achievements.
5. Come back any day. Say "let's continue my CV interview": it starts from what is missing. When you need to update your LinkedIn profile, say "update my LinkedIn".
A profile is a folder with four YAML files. Create one with honest-cv init ~/cv-profile, or let
the interview prompt fill it in with you. Keep it out of public repositories.
Any text can be a single string or one string per language: {en: ..., es: ...}. CVs can be
built in English and in Spanish.
Recording dozens of skills one question at a time is slow. A checklist is a Markdown file you mark in your editor, one letter per line:
Text after " — " is the evidence, and an evidence: line under a heading covers that section.
Lines without evidence are never saved: they come back as pending, and your assistant asks you
about them in groups. Upgrades are confirmed once for the whole file.
Checklists come in English and Spanish (--lang es): the Spanish names are synonyms of the
English ones, so a skill you mark in Spanish still matches a job post in English.
Any profession works. The built-in checklists cover software (backend, frontend, mobile, data,
platform/SRE, QA, security, ML/AI, embedded, game development, architecture, engineering
leadership) and data analysis/BI, product and project management, UX design, marketing, sales,
customer support, finance and accounting, human resources, supply chain, nursing and teaching. A template is just a
Markdown list in this format: save your own (or one someone shared) in ~/cv-profile/checklists/
and it shows up in checklist list and combines with the others in checklist new, which also
takes a path. A template can declare synonyms, such as same: Work Breakdown Structure = WBS,
and the job score, the gaps and the letter check use them, along with its skill names. Start from
examples/checklists/event-planning.md. You can also apply a file you marked yourself, in English or Spanish (años,
último uso). When you're done, honest-cv names the P and L skills that no achievement shows, and
your assistant asks you for one real example of each.
New checklists for other kinds of roles are welcome: see CONTRIBUTING.md.
See examples/ada/ for a complete, fictional profile with three variants, a job
post and its keywords.
uvx honest-cv.build, verify, score,
skill) and edit the YAML files yourself.honest-cv import reads a JSON Resume file or LinkedIn's data export
(Settings → Data privacy → Download your data, a ZIP of CSVs, or its folder if your browser
unzipped it). Ask for the full archive: the basic one, offered when you pick files, has only your
headline and summary, without roles. Contact, roles, education,
certifications, languages and each role's achievements go into your profile; with a headline and
a summary it also creates a first variant, so build works right away.
--dry-run first.imported: review their wording.imported-skills.md, a checklist for you to mark.LinkedIn's "Save to PDF" profile changes layout without notice, so it isn't parsed: with an MCP
client, the import_linkedin_pdf prompt reads its text and proposes each section for you to
approve.
honest-cv apps keeps a log in applications.yaml, inside your profile folder: which variant and
files went to which company, and each status with its date (draft, applied, interview,
offer, rejected, withdrawn). Logging the same company and role again updates it. With an MCP
client, your assistant offers to log an application after building the CV or letter, and
list_applications points out the ones with no news for two weeks. It keeps a copy of each file
you sent in sent/, since the next build overwrites out/, and the job post's text in posts/,
since links expire, and what to prepare for the interview: ask your assistant
"what did I need to prepare for Acme?".
When you get the interview, say "help me prepare the Acme interview" (or run the
interview_prep prompt). For each keyword of the saved post, honest-cv lays out your level,
your evidence and the bullets that back it, starting with the ones in the CV you sent. Your
assistant drafts the likely questions with STAR answers taken only from those bullets, honest
answers for what you only know at familiarity level or don't have, and questions to ask them. The
notes you approve are saved in prep/. Without an AI client: honest-cv apps prep ~/cv-profile acme_sre -k "Kafka,Scala".
honest-cv linkedin turns a variant into LinkedIn's texts, each within its limit: the headline
(your title and up to five production skills), About (the summary and impact bullets), one
description per position from its bullets, and the skills list. It writes them to a Markdown file
to copy from. The skills are P and L only: LinkedIn can't label familiarity, so F skills stay out,
and X skills never appear. With an MCP client, the linkedin_profile prompt polishes the wording
and checks it with check_claims.
To go the other way, honest-cv export writes a variant as JSON Resume for other CV tools and
sites, with the same rules as the CV: X skills never appear, and F skills only in a "Familiarity"
group with that level.
The cover_letter prompt drafts a one-page letter for a job post from your own bullets. Before it
is built, check_claims reads the text against your profile:
The letter uses the CV's fonts and passes the same ATS checks. Without an AI client, write it in a
text file (paragraphs separated by blank lines) and use honest-cv letter check and letter build.
Run any command with uvx honest-cv ..., or install it once with uv tool install honest-cv to
type just honest-cv.
verify works on any PDF. Use --text to see exactly what a parser extracts. Its
single_column check also catches a common defect: a date aligned to the right, which parsers
split away from its role.
The template enforces these. verify checks them on every build, and the test suite runs in CI on
Linux, macOS and Windows:
The repository's .mcp.json registers honest-cv-dev with the fictional profile, so Claude Code
in this folder can use your working copy. See CONTRIBUTING.md.
MIT. The bundled Carlito font is under the SIL Open Font License; see
src/honest_cv/render/fonts/OFL.txt.
來源：README.md，提交 51bcc81
Create a new profile folder with empty, commented YAML templates (profile, bullets, skills, variants). Use it once, before anything else, when the person has no profile; to start from LinkedIn or a JSON Resume, call import_profile afterwards. Refuses an existing profile and overwrites nothing. Returns `created`: the files written.
profile必填Read the person's skill inventory. Use it at the start of a session and before proposing skill changes; for CV content (roles, bullets, variants) use list_variants instead. Returns `levels` (what P, L, F and X mean) and `skills` grouped by level, each with its years, last_used_year, evidence, aliases and notes. Read-only.
profileAdd or change one skill in the inventory. Use it for a single skill; for several at once use set_skills, and for a whole marked list use checklist_apply. Raising a level (a new skill counts as rising from X) needs `evidence`, and the server asks the person to confirm before saving; lowering is always allowed. Returns `saved` (the stored skill) and `previous_level`.
name必填level必填notesyearsaliasesprofileevidencelast_used_yearAdd or change several skills in one call, for example a group the person just told you about. Use set_skill for one skill and checklist_apply for a marked checklist. Every upgrade needs `evidence`; the person confirms upgrades in a form, up to 10 per call, and only those they answer "sí" are saved. Returns `saved`, `upgrades_confirmed`, `upgrades_refused` and `upgrades_not_asked_yet`: when that list is not empty, call set_skills again with those skills.
skills必填profileImport an existing CV into the profile: a JSON Resume (.json), a LinkedIn data export (.zip or its unzipped folder) or a LinkedIn profile PDF. Use it at the start, right after init_profile, instead of typing everything; run it first with `dry_run`. Adds contact, roles with their achievements, education, certifications, courses, publications and languages; only what the profile lacks, never overwriting. Skills arrive without a level, so they go to <profile>/imported-skills.md for the person to mark, never to the inventory. Imported bullets are tagged `imported`: review them with the person. Returns what was added and an optional `note` to pass on. For a PDF nothing is written: it returns the `text`, to organize with the editing tools (see the import_linkedin_pdf prompt).
path必填dry_runprofileList the skill checklist templates the person can combine. Use it before checklist_new, to let the person pick the ones that fit their profession. Returns `checklists`: each with its `id`, `title`, `description`, number of `skills` and `origin` ("built-in", or "yours" for templates in <profile>/checklists/). Any profession works. Pass `lang` in the person's language to get the translated templates. Read-only.
langprofileWrite a checklist file for the person to mark in their editor, merged from the templates picked in checklist_list, without duplicates and prefilled from the inventory. Use it when the person has many skills: it is much faster than asking one by one. Next step: once they have marked it, call checklist_apply. Refuses to replace an existing file unless `overwrite`. `lists` may be empty when `extra` names the skills. Returns the file's `path` and its number of `skills`.
langpathextralists必填profileoverwriteSave the skills the person marked in a checklist file. Use it after checklist_new, once the person says they are done, and again to resume. Lines that keep or lower a level are saved at once; upgrades with evidence are confirmed in a form, up to 10 per call: call again while `upgrades_not_asked_yet` is not empty. Returns the counts saved, `pending_evidence` (marked lines without evidence: ask where they used them, in groups, and save with set_skills), `skills_without_achievements` and `key_skills_without_achievements`. Ask about nothing else: unmarked lines are for the person to mark in the file. After the checklist, ask for the person's biggest achievements first, since one usually covers several skills, then for the key skills still without one; save each with add_bullet. The rest wait for a job post that asks for them (gap_report → `no_achievement`). Never write an achievement for them.
path必填profileAdd one achievement to the bullet pool, the material every CV variant picks from. Use it after the person approves the wording, one bullet per call; to change which bullets a CV shows, use save_variant. `role` is a role id from upsert_role; leave it out for bullets used only in Selected Impact. Use only facts the person stated, never invented numbers. Returns `saved`.
id必填roletagsmetricsprofiletext_en必填text_esUpdate the person's contact details on the CV. Use it for name, email, phone, location and links; roles, education and the rest have their own tools. Only the fields given change; pass "clear" to remove an optional one. `links` replaces the whole list. Returns the `contact` as saved in profile.yaml.
nameemaillinksphoneprofilelocationReplace the certifications, courses or spoken-languages line of the profile. Use it for those three lists only; roles, education and publications have their own tools. Each list you pass replaces the current one in full; leave a field out to keep it. A course or training without an exam goes in `courses`, never in `certifications`. Returns the saved values.
coursesprofilelanguagescertificationsAdd a job role, or change some fields of an existing one by its `id`. Use it while interviewing the person about their work history; their achievements go in with add_bullet, pointing to this id. A new role needs `title`, `company` and `start`; roles stay in date order, newest first. Keep the person's real title. Returns the `role` as saved.
id必填endstarttitlecompanycontextprofilelocationAdd an education entry, or change the one with the same institution and degree. Use it for degrees and formal studies; courses and certifications go in set_profile_details. Say honestly whether the degree was completed, e.g. "B.Sc. (unfinished)". Returns the `education` entry as saved.
endstartdegreeprofileinstitution必填Add a book, article or paper the person wrote, or change some fields of the one with this `id`. Use it when the person mentions something they published; courses and certifications go in set_profile_details. Every variant shows all publications unless it lists the ones it wants (save_variant → `publications`). Returns the `publication` as saved.
id必填urlyeartitlevenueprofileCreate or replace a CV variant: the plan for one kind of role (title, summary, which bullets and skills to show). Use it to set up a variant or tailor one to a job post; read the current ones with list_variants first. It replaces the variant whole, so pass every field it should define (list_variants → `own` shows them). A variant without `extends` needs `title` and `summary`. Every reference is checked before saving. Returns `saved` and the `variant`.
id必填rolestitleimpactskillsextendsprofilesummarymax_pagescompetenciespublicationsDelete a skill from the inventory, e.g. a duplicate or a misspelling, and from the variants' skill groups and competencies. Use it only for mistakes: to record that the person doesn't have a skill, use set_skill with level X instead, which keeps gap reports honest. Ask the person first; it cannot be undone. Returns what was removed and `variants_updated`.
name必填profileDelete an achievement from the pool and from every variant that uses it. Use it when the person wants it gone; to stop showing it in one CV, change that variant with save_variant instead. Ask the person first; it cannot be undone. Returns what was removed and `variants_updated`.
id必填profileDelete a publication from the profile and from every variant that lists it. To hide it from one CV only, change that variant's `publications` with save_variant instead. Ask the person first; it cannot be undone. Returns what was removed and `variants_updated`.
id必填profileDelete a role from the profile and from every variant. A role that still has bullets is refused unless `with_bullets` is true, which deletes them too. To hide a role from one CV, edit that variant with save_variant instead. Ask the person first; it cannot be undone. Returns what was removed and `variants_updated`.
id必填profilewith_bulletsRead every CV variant, with `extends` already applied and the text of each bullet it uses. Use it before save_variant, build_cv or score_job, to know what each CV contains. Returns `variants` (with `own`: the fields each one defines itself) and `unused_bullets`: achievements no variant shows yet. Read-only.
profileRender a CV variant to PDF and/or DOCX, then verify the PDF for ATS safety. Use it when the person wants the files; for a cover letter use build_letter, and to check a PDF made elsewhere use verify_pdf. Files go to `out_dir`, default <profile>/out, replacing earlier builds of the same variant and language. Returns the `files`, each PDF with its pages, `ats_ok` and failed checks. A failed check is probably a bug in honest-cv: report it, never reword the person's text to pass it.
langformatsout_dirprofilevariant必填Check whether any PDF is ATS-safe: one column, no ligatures, dates on their role's line, text extractable, and the page count. Use it on a CV from any tool, e.g. the person's current one; build_cv already runs it on the CVs it renders. Returns `ok`, `pages`, `checks` (each with `name`, `ok`, `detail`) and `text`: what an ATS reads. Read-only.
path必填max_pagesCheck any text about the person against their profile before it goes anywhere: a cover letter, a summary, a LinkedIn text. build_letter and linkedin_profile run it too. Returns `problems`, which must be fixed (a skill marked X, or a familiarity skill not called familiarity in its sentence), and `warnings` for the person (a technology the profile doesn't show, or a number found nowhere in it). Fix them by removing or qualifying the claim, never by adding skills. Read-only.
text必填profilekeywordsRender an approved cover letter to a one-page PDF and/or DOCX with the CV's ATS checks. Use it after drafting the letter with the person and running check_claims; for the CV itself use build_cv. It runs check_claims again and refuses while there are problems. Returns the `files` and any `warnings` to show the person.
langrole必填closingcompany必填formatsout_dirprofilegreetingkeywordsrecipientparagraphs必填Export a CV variant as a JSON Resume file, for other CV tools and sites that import that format. Use build_cv for a PDF or DOCX instead. Same honesty rules as the CV: X skills never appear, and F skills only in a "Familiarity" group. Writes to `out_dir`, default <profile>/out, and returns its `path`.
langout_dirprofilevariant必填Write the LinkedIn texts for a variant, each within LinkedIn's limit: headline, About, one description per position from its bullets, and the skills (P and L only, since LinkedIn can't label familiarity). Use it when the person updates their LinkedIn profile. Writes them to a Markdown file to copy from and returns them with their `path`, plus `left_out`: bullets that didn't fit. After the person approves a polished headline or About, call it again with them: they are checked like a letter and saved in that file.
langaboutout_dirprofilevariant必填headlineRecord a job application, or update the one for the same company and role. Use it when the person sends a CV or letter, and whenever they tell you what happened (interview, offer, rejection); list_applications reads them back. Fields not given keep their value. Save the post's full text in `post`, since links expire, and your interview advice in `prep`. Each file in `files` is copied to <profile>/sent/<id>/, so the record keeps what was sent. Returns the `application` and `not_found`: paths it could not copy.
urldatelangpostpreprole必填filesnotesstatuscompany必填profilevariantList the person's job applications, most recently updated first. Use it when they ask how their search is going or which application to follow up; log_application records changes. Returns `applications` (each with its id, status and `days_since_update`), a count per status and `waiting`: open applications with no news for 14 days or more. Read-only.
statusprofilePrepare the interview for one logged application from what the person can honestly back up. Call it first with only `application` to get the saved job post (`post_text`), then again with the post's keywords. For each keyword it returns their level, evidence, `stories` (the bullets that show it, those in the CV they sent first) and a `standing`: strong (tell a story from those bullets), no_achievement (ask for a real example), marked_pending (ask for evidence), familiarity (say it is familiarity) or lacking (say so honestly). Every story must come from `stories`. With `notes`, saves the approved prep to <profile>/prep/<id>.md.
notesprofilekeywordspreferredapplication必填Measure how well each CV variant covers a job post's keywords, best first. Use it first when the person shares a post, to pick a variant; then call gap_report on that variant to see what is missing and why. Extract the keywords yourself in the post's own wording: `keywords` are the required ones, `preferred` the nice-to-haves. Returns `scores`: per variant, the required and overall coverage and each keyword found, found only as familiarity, or missing. Read-only.
langprofilekeywords必填variantspreferredExplain the keywords a job post asks for that one variant doesn't show, grouped by what can honestly be done. Use it after score_job picked the variant, before editing it with save_variant. Returns: `in_inventory_not_in_cv` (the person has it: add it, using `unused_bullets`, or name it in the summary or competencies), `familiarity_only` (only as familiarity), `not_in_inventory` (do not claim it), `no_achievement` (P or L skills no bullet shows: ask for one real achievement each) and `marked_in_checklist` (marked in their checklist, still without evidence: ask where they used it). Read-only.
langprofilevariant必填keywords必填preferred已滿足適用檢查項的 90%
該評分衡量此收錄條目的完整程度，依據註冊中心條目、端點與公開來源計算得出。它不是對伺服器本身的評價，也不表示伺服器安全或目前可用：可連通性是帶日期的檢查結果，評分也不包含安裝量或使用者評價。
請在註冊中心條目中宣告憑證請求標頭，或保持端點能夠回應探測，以便判定其認證方式。
claude mcp add honest-cv -- uvx honest-cv serve --profile ~/cv-profilemkdir -p ~/.claude/skills/honest-cvcurl -fsSL https://raw.githubusercontent.com/JesusCabreraReveles/honest-cv/master/skills/honest-cv/SKILL.md \ -o ~/.claude/skills/honest-cv/SKILL.md{ "mcpServers": { "honest-cv": { "command": "uvx", "args": ["honest-cv", "serve", "--profile", "/absolute/path/to/cv-profile"] } }}curl -LsSf https://astral.sh/uv/install.sh | sh # Windows: see docs.astral.sh/uvclaude mcp add honest-cv -- uvx honest-cv serve --profile ~/cv-profile## Messaging and streamingevidence: Payments platform at Acme, 2020–2026- [P] Kafka — reconciliation flow, 2M events a day · years: 6 · last year: 2026- [F] RabbitMQ — a workshop and the official tutorials- [X] NATShonest-cv checklist list # backend-jvm, platform-sre, ...honest-cv checklist new ~/cv-profile backend-jvm engineering-leadhonest-cv checklist apply ~/cv-profile ~/cv-profile/checklist.mdhonest-cv checklist list ~/cv-profile # yours too, e.g. accountinghonest-cv checklist new ~/cv-profile accounting engineering-leadhonest-cv init ~/cv-profile # create the profile templateshonest-cv build ~/cv-profile -v backend -l en -l es # PDF + DOCX, verified on every buildhonest-cv verify any-cv.pdf --text # check any PDF, even one made elsewherehonest-cv score ~/cv-profile -k "Kubernetes,Go" -p "Rust" # coverage per variant, with gapshonest-cv skill ~/cv-profile "Rust" L --evidence "CLI tool on GitHub" # asks you to confirmhonest-cv import ~/cv-profile linkedin-export.zip --dry-run # or resume.jsonhonest-cv export ~/cv-profile -v backend -l en # JSON Resumehonest-cv apps log ~/cv-profile --company Acme --role "SRE" -s applied -v platformhonest-cv apps list ~/cv-profile # latest first, with countshonest-cv apps prep ~/cv-profile acme_sre -k "Kafka,Scala" # what backs each keyword, for the interviewhonest-cv linkedin ~/cv-profile -v platform -l en # texts to paste into LinkedInhonest-cv remove skill ~/cv-profile "Kafka" # asks you to confirmhonest-cv letter check ~/cv-profile letter.txt -k "Kubernetes,Go" # claims your profile doesn't supporthonest-cv letter build ~/cv-profile letter.txt --company Acme --role "Platform Engineer" -l enhonest-cv serve --profile ~/cv-profile # the MCP server, over stdiouv syncuv run pytestuv run honest-cv build examples/ada -l en -l es