agentleFS
rafaym1rafaym1/scholarship-desk/.claude/skills/tailor/SKILL.md

tailor

Skill18 starsChanged 4 days agoMITView on GitHub

Prepare one application's written material — read the call and its questions, pick the best of the applicant's CVs, draft every answer from the fact bank in the applicant's own voice within limits, run checks, and send drafts for approval in the dashboard. Use "/tailor <opportunity id>".

10 other projects use a version of this.

Adopted by

Most recently taken up by

Used with

Projects with it also use inbox.

Other files people use alongside this one.

Files found in the same public projects as a version of this one, more often than they turn up anywhere else. Each project counts once, large collections count for less, and this file's own repository is left out.
  • lamb92009Skill

    inbox

    Use when the user runs /inbox or asks to process the inbox folder in the RDL vault. Reads unprocessed files, routes content to correct vault locations, creates or updates notes with wikilinks, and deletes processed originals.

    Also in 3 projects that use this

The file

We split this skill into the rules it gives an agent

6 rules. 0 are common, 4 are distinctive, and 2 are very specific to this project.

  • 0

    Common

    Most of the 8 edited versions we read keep these.

  • 4

    Distinctive

    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
  1. Distinctive:

    Read CLAUDE.md first. Opus gives the best drafts.

    line 8
  2. Very specific:

    If the opportunity is not eligible/needs_confirmation, say why and stop, unless the applicant says to apply anyway: python -m scholar app …

    line 23
  3. Distinctive:

    AI policy: look for rules on AI-assisted writing in the call/FAQ.

    line 32
  4. Distinctive:

    Start from approved answers of the same category if any exist (prefer ones tied to good outcomes), then adapt to this funder's criteria and …

    line 49
  5. Very specific:

    Use only the fact bank and the narrative.

    line 51
  6. Distinctive:

    Write in the applicant's voice (voice sample).

    line 56

“Common” is measured: kept by at least 6 in 10 of the 8 edited versions of this file we could read. “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

Tailor an application

Read CLAUDE.md first. Opus gives the best drafts; Sonnet is fine for revisions.

Gather

python -m scholar opp show <id>
python -m scholar drafts show <id>            # existing brief, drafts, your feedback
python -m scholar library cv-cards
python -m scholar library facts
python -m scholar library show-voice
python -m scholar library answers --tag <category>   # answers approved before, by question type
python -m scholar app stats                   # which CVs led to interviews/wins
python -m scholar profile show --section narrative

If the opportunity is not eligible/needs_confirmation, say why and stop, unless the applicant says to apply anyway: python -m scholar app override <id> --reason "<their words>". Then python -m scholar app prepare <id>.

1. Brief

  • Questions: from the record (portal recon) or the official call. If the portal hasn't been walked and questions aren't published, run /recon <id> first.
  • Criteria: what this funder scores (selection criteria, priorities) from the official page.
  • AI policy: look for rules on AI-assisted writing in the call/FAQ; record the wording or "not stated". If AI help is prohibited, tell the applicant before drafting and let them decide.
  • CV choice: compare the CV cards against the criteria and format rules (page limit, Europass, academic vs industry). Pick one, give a one-line reason and the runner-up. If none fits a required format, propose a tailored version (reorder/trim/re-emphasise only, no new facts); creating that file needs the applicant's OK, then profile doc-add cv ... and /ingest.

Write data/tmp/brief.json:

{"criteria": ["Leadership", "Networking", "Return-home impact"],
 "cv_choice": {"path": "documents/CV - leadership.pdf", "reason": "...", "runner_up": {"path": "documents/CV - research.pdf", "reason": "..."}},
 "ai_policy": "not stated", "questions": null}

python -m scholar drafts brief <id> data/tmp/brief.json (questions: null uses the record's questions).

2. Draft each answer

  • Start from approved answers of the same category if any exist (prefer ones tied to good outcomes), then adapt to this funder's criteria and wording.
  • Use only the fact bank and the narrative. Every claim must trace to a fact id — pass them with --sources.
  • Facts tagged fellowship come from fellowship essays: for degree and PhD applications use them only when nothing else covers the point. Facts tagged sensitive_disclosure are used only if the applicant has said they may be (their answer is saved as a fact with the tag preference); if there is no such fact, ask before drafting.
  • Write in the applicant's voice (voice sample). Specific and plain; no clichés they didn't use; no invented numbers, names, events or feelings.
  • Answer exactly what is asked and name the correct programme/sponsor.
  • Aim for 90–100% of the limit; never exceed it.
  • Missing information (plans after the degree, why this country, a specific story): ask the applicant in one consolidated list, then save the answers as facts (library add-facts with "source": "applicant, <date>") before using them.

Save each answer: write data/tmp/<id>_<qid>.md → python -m scholar drafts put <id> <qid> data/tmp/<id>_<qid>.md --sources f3,f7. Fix anything listed under blocking (over limit, placeholders, another scholarship's name).

3. Review pass

Re-read every draft against its question and its source facts: does it answer the question? Is every claim supported? Any wrong programme name? Fix and drafts put again.

4. Approval

Tell the applicant the drafts are in the dashboard → Drafts tab, where they can approve, edit or request changes. If they approve a specific answer in chat instead, run python -m scholar drafts approve <id> <qid> for that answer only. For "changes requested", read their feedback in drafts show <id> and revise only those answers.

5. Documents and readiness

  • python -m scholar docs check <id> — fix what it reports (merge PDFs: python -m scholar docs merge "documents/merged.pdf" "documents/a.pdf" "documents/b.pdf", then profile doc-add; ask before creating files).
  • Referees: prepare a short brief they can use (funder criteria, chosen CV, 3–4 points worth highlighting). Never write a letter in a referee's voice.
  • python -m scholar app readiness <id> — report remaining blockers. When ready, the next step is /apply <id>.

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.