agentleFS
HimadriTryingHimadriTrying/ai-job-search-agent/.claude/skills/cover-letter/SKILL.md

cover-letter

Skill1 starsChanged 3 months agoMITView on GitHub

Draft a targeted cover letter for one role, in the user's voice, grounded only in real facts. Requires company research, runs the style checker, the Reviewer and the honesty gate. On demand.

36 other projects use a version of this.

Adopted by

Most recently taken up by

Used with

Projects with it also use interview-prep, tailor-resume and 6 more.

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.
  • AnthropicSkill

    interview-prep

    Create structured interview plans with competency-based questions and scorecards. Trigger with "interview plan for", "interview questions for", "how should we interview", "scorecard for", or when the user is preparing to interview candidates.

    Also in 8 projects that use this

  • proficientlyjobsSkill

    tailor-resume

    Tailor your resume for a specific job posting

    Also in 4 projects that use this

  • rhowardstoneSkill

    job-search

    Search for jobs across multiple boards and company career pages. Use when looking for new positions to apply to.

    Also in 4 projects that use this

  • aslavchevSkill

    mock-interview

    Simulate a real job interview for any role. Use when user says "mock interview", "interview me", "interview prep", "practice interview", or wants to rehearse for a specific company or position. Adapts rounds and questions to the role — technical, managerial, creative, or any other.

    Also in 3 projects that use this

  • rhowardstoneSkill

    apply

    Research-backed job application workflow emphasizing relationship-building over mass applications

    Also in 3 projects that use this

  • JDerekLomasSkill

    literature-review

    Conduct comprehensive, systematic literature reviews using multiple academic databases (PubMed, arXiv, bioRxiv, Semantic Scholar, etc.). This skill should be used when conducting systematic literature reviews, meta-analyses, research synthesis, or comprehensive literature searches across biomedical, scientific, and technical domains. Creates professionally formatted markdown documents and PDFs with verified citations in multiple citation styles (APA, Nature, Vancouver, etc.).

    Also in 3 projects that use this

The file

We split this skill into the rules it gives an agent

3 rules. 0 are common, 0 are distinctive, and 3 are very specific to this project.

  • 0

    Common

    Most of the 32 edited versions we read keep these.

  • 0

    Distinctive

    This author's own choices. Worth a look.

  • 3

    Very specific

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

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

    File what it found: before moving on, or the turn will not end: python gates/session.py record --event draft-reviewed --path <draft> --note …

    line 50
  2. Very specific:

    Honesty gate: must pass: python honesty/verify.py <draft> --target "<Company>" --job <jd-file> — the flags tell the gate which company/JD …

    line 52
  3. Very specific:

    If they push back on the draft, run learn: Ask whether the correction is a one-off or a standing rule, and store a standing one in the …

    line 57

“Common” is measured: kept by at least 6 in 10 of the 32 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

cover-letter

Precedent

Drafts a targeted cover letter from your profile and the specific JD, then runs the style checker, the Reviewer and the honesty gate.

Inputs

profile/06-cover-letter-notes.md (the house spec: read it in full before drafting), profile/03-writing-style.md, career_facts.yaml, the JD as a file (see CLAUDE.md → "Getting the JD": unfetchable URL → ask the user to paste the text into data/jd/<company>-<role>.md; never reconstruct a JD from search snippets), and the company-research output, which is required, not optional (see step 0).

Flow

  1. Research first, or stop. The letter's opening beat is a researched observation, so data/research/<company>.md (or <company>/company-research.md) must exist and be non-empty. If it does not, run company-research first, or tell the user the letter is not ready to draft. Do not substitute the model's own memory of the company: a letter that cites a company fact nobody researched will pass every style check and still be wrong, which is exactly the failure this step exists to prevent. 0b. Read the rules the user has already taught you. Before drafting a word: python scripts/learned_rules.py brief --scope letter. These are corrections they made on earlier drafts and asked to stand. Honouring them here is the whole point of having stored them; making the same correction twice is what the store exists to stop.
  2. Pick the spine, not the angle. One thing the research surfaced that your own experience answers. It must relate to the role or the domain, and it must be something only real research would find. Check it against the four tests in the notes file.
  3. Draft as one story in four beats, per the notes file: observation, flagship proof, turn, callback. In the user's voice (writing-style file). 200 to 280 words.
  4. Run the style checker and fix everything it reports: bash scripts/check-cover-letter.sh <draft.md> --company "<Company>" Its last step re-checks the user's own learned rules, so a stored correction fails the letter rather than being quietly reintroduced.
  5. Reviewer subagent (fresh context) critiques against profile/06-cover-letter-notes.md, which it should be given along with the draft and nothing else. It is looking for what the checker cannot see: two flagships where the spec allows one, a closer that would be true of a competitor, an opening that explains the company to itself, a win dressed as a constraint. Revise. If the subagent cannot be spawned, stop and tell the user. The drafter never reviews its own work as a substitute; in real use, a draft that passed all eight mechanical checks came back from the Reviewer with ten violations of the spec its own author had written. File what it found before moving on, or the turn will not end: python gates/session.py record --event draft-reviewed --path <draft> --note "<findings>".
  6. Honesty gate must pass: python honesty/verify.py <draft> --target "<Company>" --job <jd-file> — the flags tell the gate which company/JD the letter legitimately references, so only true fabrications fail. If PDF: render, check one page, signature visible, fonts match.
  7. Present. Sending stays the human's job.
  8. If they push back on the draft, run learn. Ask whether the correction is a one-off or a standing rule, and store a standing one in the same session. A correction that only fixes this letter arrives again on the next one.

What passing means

The checker is a floor, not a verdict. It proves the letter is the right length, names the right company, invents no numbers, and is structurally a story rather than a list. It cannot tell you the argument is any good. Steps 4 and 5 are what make it safe to send; step 3 only makes it worth their time to read. See docs/FAILURE-MODES.md.

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.