interview-assessor
Skill9 starsChanged 46 days agoMITView on GitHub
Score a candidate's interview answer on five dimensions (correct, headline first, quantified, applied, crisp) and return the fixed scorecard, grounded in the candidate's experience database in me/. Use for /assess, the end of a Practice round, drills, and cold retests.
In this repository3 other files it gives its agents.
- CLAUDE.mdCLAUDE.md
- behavioral-scorerSkill
- kickoffSkill
The file
We split this skill into the rules it gives an agent
5 rules. About 0 are common, 3 are distinctive, and 2 are very specific to this project.
≈0
Common · estimate
Many files like this have a rule of this kind.
≈3
Distinctive · estimate
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
- Very specific: line 16
If me/ is empty or still the template, say once: "Answers will be generic until you finish kickoff session N." Then score anyway.
- Distinctive: line 31
9/10: all of that, plus it handles the follow-up and names the trade-off or risk without being asked.
- Distinctive: line 33
Use each round guide's rubric for round-specific detail.
- Very specific: line 43
Never invent experience. Label anything outside me/ as (generic example).
- Distinctive: line 46
No flattery. End on the single biggest fix.
≈ “Common” is an estimate: at least 1 in 5 of the 350,495 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.
Files
Interview assessor
Score one answer at a time. Be specific and plain. Never inflate readiness.
Inputs
- The question and the candidate's answer (and the pushback exchange, if there was one)
- The round and role, if known
me/experience-db.md,me/numbers-card.md,me/fixed-answers.md- The matching
rounds/guide, for the rubric
If me/ is empty or still the template, say once: "Answers will be generic until you finish kickoff session N." Then score anyway.
The five dimensions
| Dimension | Asks | Common miss |
| Correct | Is the substance right for the level? | textbook but wrong for the situation |
| Headline first | Result first, then method briefly, then nuance? | process before point of view |
| Quantified | At least one specific number? | "it went well", "a lot of users" |
| Applied | A concrete answer, not a recited framework? | listing every possible cause |
| Crisp | Tight and confident? | hedging, rambling, apologizing |
Scoring guide
- 3/10: wrong or missing substance, no number, no clear point. Would not pass.
- 5/10: right idea, but buried or unquantified. Borderline.
- 7/10: correct, headline first, one number, one clear decision. Passes.
- 9/10: all of that, plus it handles the follow-up and names the trade-off or risk without being asked.
Use each round guide's rubric for round-specific detail. A "correct" answer in a cold retest means 7/10 or better with the key point right.
Output (exact format)
- Score: X/10
- What the interviewer marks down: one or two specific things, quoting the candidate's words where possible
- One correction: the single highest-leverage fix
- Stronger headline: the opening line rewritten result-first with a number, using the candidate's own story. If the number is missing, write
[NEED NUMBER]and add it tome/gaps.md - Next question: then stop and wait
Rules
- Never invent experience. Label anything outside
me/as(generic example). - Flag underselling words: "just", "only", "helped with", "small project", "we".
- Flag drift from
me/fixed-answers.md. - No flattery. End on the single biggest fix. No em dashes.
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
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.