# Interview Debrief

> Use when debriefing immediately after a real or mock interview, or when reviewing interview notes, transcripts, questions, answers, interviewer reactions, feedback, and outcomes.

- Skill: `zhanlincui/interview-debrief` (Agent Skill, multi-file: 3 files)
- Install (CLI): `npx skillmds@latest add zhanlincui/interview-debrief`
- Raw SKILL.md: https://api.skillmd.com/api/skills/zhanlincui/interview-debrief/raw
- Safety review: pending
- Works with: Claude Code, Claude.ai, OpenAI Codex
- Category: Coding & Dev Tools
- Author: ZhanlinCui (https://skillmd.com/u/zhanlincui)
- Updated: 2026-09-17
- Page: https://skillmd.com/skills/zhanlincui/interview-debrief

---


# Interview Debrief

**REQUIRED BACKGROUND:** Use `career-state-protocol` for every new fact and write-back.

## Goal

Preserve fragile interview memory, separate observation from interpretation, diagnose performance against the actual target, and improve the next round and durable Career State.

## Memory capture first

Do not analyze first. Capture fresh evidence before offering a verdict.

Ask the user to reconstruct the interview in chronological order:

1. questions and transitions;
2. the answer or story used;
3. follow-ups, interruptions, requests for detail, or topic changes;
4. interviewer reactions such as notes, pauses, challenge, agreement, or time pressure;
5. the user's strongest and weakest moments;
6. explicit feedback, next steps, and known outcome.

Ask one memory prompt at a time unless the user requests fast mode or provides a transcript. Mark recalled wording as approximate when necessary.

Read `references/capture-and-analysis.md` before analyzing and `references/debrief-contract.md` before writing results.

## Analysis

1. Reconstruct what each question was likely testing using the `T###` model.
2. Compare the answer with confirmed claims, the submitted resume, and prepared story boundaries.
3. Separate observed interviewer reactions from Target inference about what they may mean.
4. Identify answers that landed, concerns that intensified, missed opportunities, and recovery quality.
5. Update story strength, follow-up risks, and story use count.
6. Add new questions to the target question map and detect repeated patterns.
7. Convert new candidate facts mentioned in the interview into `candidate` claims pending confirmation.
8. Produce next-round priorities and the smallest useful drill plan.

## Write-back rules

Write back interview evidence to `I###`, target concerns to `T###`, story usage to `S###`, and gaps to `G###`. A transcript does not automatically verify a claim about the candidate's past; it proves only that the claim was said. New history still needs confirmation or source support.

If an outcome is unknown, do not predict it confidently. Give an evidence-based signal range and name the uncertainty.

## Output

Return the reconstructed timeline, interviewer read, answer-level findings, global patterns, next-round plan, and explicit write-back summary. Focus on the three changes with the highest expected impact.

## Boundary

Debrief ends at preparation improvement. Do not enter offer evaluation, compensation, or application follow-up workflows.