agentleFS
VincenzoImpVincenzoImp/openings/skills/prepare-applications/SKILL.md

prepare-applications

Skill7 starsChanged 3 days agoMITView on GitHub

Prepare the CV and cover letter of Openings applications from the owner's career entries, templates and Writing rules - "prepare CV and letter for every shortlisted job", "prepare the application for X", "redo the letter for Y". Compiles PDFs locally, uploads them and leaves each application Ready for the owner to send.

The file

We split this skill into the rules it gives an agent

9 rules. About 0 are common, 5 are distinctive, and 4 are very specific to this project.

  • ≈0

    Common · estimate

    Many files like this have a rule of this kind.

  • ≈5

    Distinctive · estimate

    This author's own choices. Worth a look.

  • 4

    Very specific

    Tied to their project's names, files or tools. Probably skip.

  • Common
  • Distinctive
  • Very specific
  1. Very specific:

    Read openings first, and writing before you write.

    line 11
  2. Very specific:

    Check your tools: a shell to upload, and pdflatex (or the engine the template names) for TeX templates.

    line 25
  3. Very specific:

    Read it with applications_get (up to 20 at a time) and its job with jobs_get.

    line 32
  4. Distinctive:

    Employer requirements: Before writing, look in the posting for:

    line 35
  5. Distinctive:

    every link works and points where it should.

    line 64
  6. Distinctive:

    no placeholder, TODO or other company's name left.

    line 65
  7. Distinctive:

    every claim traces to the record.

    line 66
  8. Very specific:

    Store: Call documents_create with the PDF's name and size, the TeX source when you compiled one (inline, or as a file with its sha256), the …

    line 71
  9. Distinctive:

    Read the applications again: Ready is computed from complete documents and open requirements, and each application says what is still …

    line 89

≈ “Common” is an estimate: at least 1 in 5 of the 372,429 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.

Keep what fits and drop the rest. Your version, ready for your agents.
Files
SKILL.md

Prepare CVs and letters

Read openings first, and writing before you write. You prepare; the owner reviews, sends and marks Applied.

1. Gather

  • workspace_context with the full career and templates: Writing rules, Facts about me, approved highlights (★ always include) and each template's source, guidance and limits.
  • The work: applications_list of Shortlisted applications missing a document, or the ones the owner named. Order them by deadline, then by fit. A job the owner names that has no open application gets one with applications_create; a Closed one is started again only when they ask. Skip applications that are already Ready unless asked to redo them.
  • Check your tools: a shell to upload, and pdflatex (or the engine the template names) for TeX templates. Without a compiler, a PDF-only revision from another tool is fine when the owner has no TeX template; never upload an unchecked PDF.

2. Per application

Read it with applications_get (up to 20 at a time) and its job with jobs_get. Follow its pinned note first: it is the owner's guidance.

Employer requirements. Before writing, look in the posting for:

  • a ban on AI-written applications: stop, add it as a blocking requirement and tell the owner;
  • attachments, references, transcripts, work samples: list those the owner must supply as requirements; never invent a referee;
  • format: page limit, language, CV only, form questions; when no letter is wanted, require the CV alone;
  • eligibility (work permit, degree, location) compared with Facts about me without inferring rights; an open question is a requirement, not a guess;
  • a closing date: set the deadline; a closed posting is reported, not prepared.

Record them with applications_update and the version you read. A blocking one keeps it from Ready until the owner resolves it.

Select. Pick the three to five needs of the role, then the career entries and approved highlights that answer them best. Start from ★ highlights unless the job gives a reason to drop one. Keep each item under its real role and dates.

Write. Fill the template following its guidance, the Writing rules for that document and the writing reference. Keep the template's fixed parts. The letter addresses this employer and role; the CV and letter tell one story with the same facts.

Compile and check. Compile locally, then read the PDF itself:

  • page count and template limits, nothing cut off or overflowing;
  • every link works and points where it should;
  • no placeholder, TODO or other company's name left;
  • every claim traces to the record; nothing contradicts Facts about me;
  • each Writing rule holds.

Fix and compile again until it passes.

Store. Call documents_create with the PDF's name and size, the TeX source when you compiled one (inline, or as a file with its sha256), the highlights and entries you used and, for a redo, the revision it replaces. PUT a source file first, then the PDF, with the returned grants (see openings). A revision is complete and final once its PDF is stored: check before uploading, and make a new revision for any change.

Form questions. Record the employer's form questions as the application's questions and draft answers from the record, never inventing; a long form can be an answers revision with documents_create. The owner submits them.

Gaps. What the record cannot support goes in an unpinned note on the application for the owner, never into the documents.

3. Confirm Ready

Read the applications again: Ready is computed from complete documents and open requirements, and each application says what is still missing. Fix what you can; ask about the rest. Do not set any stage: Ready applications wait for the owner.

Report back

In chat: the applications now Ready, by title and company, with any deadline; those not Ready with the reason (a requirement for the owner, a missing fact, a failed compile); the gaps you noted; and a reminder that the owner reviews, sends and marks each one Applied.

Download now (ZIP)

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
What did you use it for? e.g. Reviewing pull requests in a Rails app

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.