Skip to content
Back to skills

Interview Prep

ASecurity

Builds an interview prep sheet for one company and stage from the user's own STAR stories, profile and the role's posting, plus research on the company and the interviewers, with questions to ask. Use for /job-tracker:prep, when an interview is booked or a booking link arrives, or when the user asks how to prepare for a specific interview.

  • 4 stars
  • 0 votes
  • 0 copies
  • 0 views
  • Added November 18, 2026
ai-agentsapi

Works with

  • api

Security analysis

A100/100

Scanned November 18, 2026

npx -y skills add adamwardecki/job-tracker-mcp --skill interview-prep --agent claude-code

Installs into .claude/skills of the current project.

Are you the author of Interview Prep?

Add the live security badge to your README. It updates with every re-scan.

Security grade badge for Interview Prep
[![Security: A — Skills Directory](https://www.skillsdirectory.com/api/skills/adamwardecki-interview-prep/badge)](https://www.skillsdirectory.com/skills/adamwardecki-interview-prep)

More formats (shields.io, HTML) on the badges page. Keep it an A: scan every change in CI with Pro.

Download with Pro
SKILL.md
---
name: interview-prep
description: Builds an interview prep sheet for one company and stage from the user's own STAR stories, profile and the role's posting, plus research on the company and the interviewers, with questions to ask. Use for /job-tracker:prep, when an interview is booked or a booking link arrives, or when the user asks how to prepare for a specific interview.
---

# Interview prep

Help the user walk in with the right stories ready. Everything they will say must be true and theirs.

## Hard rules

- **No invented claims.** Every story, number and technology in the prep comes from the `stories` or `profile` document (or `reusable-answers`). If a likely question has no honest story behind it, say so to the user in a "Thin spots" section with an honest way to answer. Never draft a fictional example.
- **Use the user's words.** Adapt their stories; do not rewrite them into a different voice or add drama.
- **Research is cited.** Each company fact carries its source (URL). Mark anything older than a year.
- **Stage-aware.** A recruiter screen, a hiring-manager call, a system design round and a final with a founder need different prep. Ask the stage if the tracker does not have it.
- **Short.** One screen per section. The user reads this before a call, not instead of preparing.
- Interviewer research is limited to professional, public information (role, team, public talks or posts about the work). Nothing personal.

## Inputs

- `find_company` and `open_items`: the role, the interview (`starts_at`, `tz`, `stage`, `people`, `link`, `notes`).
- The role's posting (employer page), `score_notes`, the cover note and form answers sent (artifacts), and any recruiter emails in the thread.
- `get_document`: `stories`, `profile`, `reusable-answers`.
- Rules: load `inbox`, plus `writing` if you draft any text for the employer (Reading rules: `${CLAUDE_PLUGIN_ROOT}/templates/rules.md`; Hermes: `${HERMES_SKILL_DIR}/../../templates/rules.md`).
- Web search and fetch: the company site, product docs, changelog or blog, recent funding or launches, engineering blog, public talks by the interviewers.

## Procedure

1. **Pin the interview.** Company, role, stage, time in the user's timezone, people, format (video, take-home, live coding, system design, case, portfolio review, work sample, panel, teaching demo).
2. **Re-read what was sent.** The cover note and answers the employer has seen. The interview will probe them; the prep must be consistent with them.
3. **Research the company** (10 to 15 minutes of reading, not more):
   - What the product does, for whom, and how it makes money.
   - Stage, size, funding, recent launches.
   - How the work is done there: for engineering, stack, how they ship and public incidents or write-ups; for other fields, the tools, methods, standards or regulators and public work that apply.
   - The interviewers' roles and what they have written or said publicly about the work.
4. **Predict the questions.** From the posting's requirements and responsibilities, the stage, and the company's current problems. 8 to 12 questions, most likely first.
5. **Map stories to questions** with STAR. For each predicted question pick the best story from `stories`, and write:
   - **Situation** (one line), **Task** (one line), **Action** (the user's own actions, 2 to 3 lines), **Result** (with the number if the story has one).
   - Which requirement in the posting it answers.
   - The likely follow-up and the honest short answer.
   A story can serve several questions; note which angle to lead with for each.
6. **Thin spots.** Requirements with no strong story. For each: the closest true experience, and a plain honest answer if asked directly. Gaps are answered honestly when asked; they are not raised unprompted.
7. **Questions to ask** (5 to 7), specific to this company and stage. Prefer questions about how the work happens (how decisions get made, what the first 90 days look like, how they ship and measure, what is hard right now) over things the website answers. Include one about whatever made the user want the role.
8. **Logistics.** Time in the user's timezone, link, who is on the call, what to have open (CV, the posting, the cover note), and the booking state on the invite calendar.
9. Save the sheet with `add_artifact` (`kind: "note"`, `ref: "prep:<stage>"`) and show it.

## Sheet layout

```
# <Company>, <Role>: <stage>, <day time tz>
People: ... | Link: ... | Format: ...

## The company in five lines
## Why you (two lines, from the cover note)
## Likely questions → stories (STAR)
## Thin spots and honest answers
## Questions to ask
## Logistics
```

## System design and technical rounds

- Pull the user's real systems from `profile` and `stories` that resemble the company's product, and list the trade-offs they actually made.
- Suggest a short practice outline (requirements, API, data model, scaling, failure modes) using the company's own domain as the example.
- Do not invent architecture the user did not build.

Attribution

Is this your skill, or is something wrong with this listing? Request removal or report an issue. Author removals are honored within 72 hours.

View all in ai-agents →

Comments

Loading comments…

Newsletter

Get the weekly list of safe new skills, and the ones we pulled.