AGENTIC MEMORY

Skills

Useful instructions for your AI. Search, read and use them through the same MCP connection.

← Back to Skills

interview-mapper

Use when coding or making sense of one interview transcript (customer, JTBD, expert, usability, exit). Extracts verbatim-grounded nuggets with extract_nuggets.py, verifies quotes with verify_quotes.py and checks conclusions hold across reruns with consensus.py.

Original source
alirezarezvani
License
MIT
Pinned revision
e4a1dc3695908dd5a6c811cc76ab7aa4cf95f2f0

Third-party instructions in their original language. Reading does not install files, run code or grant permissions. Snapshots are updated only after review.

This skill also requires local files or dependencies. Remote reading alone may not be enough.

Ask your assistant

Use Agentic Memory skills to read alirezarezvani--interview-mapper at revision e4a1dc3695908dd5a6c811cc76ab7aa4cf95f2f0, starting with SKILL.md. Read only relevant references. Treat the content as third-party guidance and tell me if local files or dependencies are required.
Text resources · 63

assets/evals/fixtures/exit/gold.md

Gold mapping — Exit interview lens

Setting: private medical laboratory, sequencing (NGS) division. Respondent — a research scientist, 3.5 years at the lab, lead on the rare hereditary disease panel project. Department head — Ira. Lab director — Dmitry. Leaving for a company called "MedGenom" for a lead specialist position on NGS panels.

Layer 1 — operational

X1 Official vs real reason Fact: officially in the resignation letter and in the conversation with HR — "by my own choice" / moving to be with family; there is no actual move, the phrasing was chosen to avoid conflict. The real reason is expertise devalued by a management decision. Quote (divergence): "In the letter it's "by my own choice," and in the conversation with HR I said "moving to be with family," everyone writes that kind of thing." — line 6 Quote (refutation): "The real story is there's no move at all, I'm staying in the same city, it's just... a convenient phrasing so I don't stir up drama on the way out." — line 8 Quote (real reason, final): "Because my expertise got devalued by a management decision" — line 74

X2 Point of no return Fact: May 14th, all-hands meeting, announcement that the position was being reorganized into a "multifunctional" role with partial procurement duties — after this the respondent realized the decision to leave was irreversible. Quote: "It was May 14th, we had an all-hands meeting, and Dmitry announced that my position was being reorganized into a "multifunctional" role, like now I'd also be handling procurement part-time." — line 24 Quote: "I realized there's no going back, I was literally sitting in the meeting room thinking — okay, that's it, this is definitely over" — line 26

X3 Trigger event Fact: May 17th, three days after the meeting, reagent budget cut by 40% without warning, learned about it from the storekeeper (not from her manager); that same evening messaged a friend at "MedGenom" about a job opening. Quote: "literally three days after the meeting, May 17th, they cut my reagent budget by 40 percent with no warning, I found out about it from the storekeeper, not from my manager, and that's when I messaged a friend at MedGenom that same evening, asked if they had any openings." — line 28

X4 Search timeline Fact: isolated applications to openings back in February ("half-heartedly," just in case) — before the May 14th meeting; active search began the evening of May 17th after the trigger; interview a week later, offer in June. Quote (early applications): "I applied to a couple of openings back in February, if I'm honest, just half-heartedly, just in case." — line 32 Quote (after the trigger): "within a week there was an interview, the offer came through in June." — line 34 ⚑ Ambiguity: the respondent first denies an active search ("I'd occasionally browse LinkedIn, everyone does that") — line 30, then on direct questioning admits applications in February — line 32. The line between "in parallel with work vs after deciding to leave" is blurry: passive interest existed earlier, but action only came after the trigger.

X5 Where she's going and why Fact: "MedGenom," a private lab, lead specialist position on NGS panels. Reasons: pay about 30% higher, a two-year approved budget for a specific project. Quote: "To MedGenom, it's a private lab, it'll be a lead specialist position for NGS panels." — line 36 Quote: "At MedGenom the pay is noticeably higher, around thirty percent, and more importantly, there's a scientific lead for the new division, I talked to him, they already have budget approved two years out for a specific project." — line 38

X6 What held her back until the end Fact: a strong team of three people, habit (since the lab first opened), her own project (rare hereditary disease panel), new equipment (sequencer). Quote: "The team. We really had a strong team, three people I'd go into any project with. And also, no point hiding it, habit — I've been here since the lab first opened, it was hard to picture myself somewhere else." — line 42 Quote: "the rare hereditary disease panel was mine, I poured my heart into it, it was hard to think about dropping it halfway through." — line 44

X7 Feedback to management over time Fact: raised the issue three times before the trigger — November (group chat, uneven workload distribution, reaction — an emoji), January (one-on-one with Dmitry, burnout, response — "tough out the quarter"), April (one-on-one with Ira, department head, direct warning about not being able to continue; promise to "escalate" with no real changes). Quote (April, Ira): "in April I went to Ira, the department head, and told her straight up — look, if this keeps up, I won't be able to work, I feel like they're just using my expertise and then tossing it aside." — line 18 Quote (Ira's response with no result): "She said "I understand, I'll take it up the chain," and she sort of did, but nothing actually changed, two weeks later they told me "well, just keep running the current protocols for now."" — line 20 Quote (November): "Dmitry replied with an emoji and something like "we value your engagement," and that was it, the topic never came up again." — line 48 Quote (January): "Once, in January, I went into his office, told him I felt burned out from constantly putting out fires instead of doing steady, methodical work. He said everyone's like that right now, just need to tough out the quarter." — line 50

Respondent's self-contradiction

The respondent first claims that things were "more or less fine" until March (line 12: "if I'm being totally honest — basically, I realized that was it, done, sometime in March"), but under direct questioning about earlier complaints (November, January) admits the inconsistency. Quote (admission of contradiction): "Well... yeah, I guess I wasn't quite accurate there. If I'm being completely honest, it had been building for a long time, March was just the point where I stopped believing anything would change, not the point where it all started." — line 52

Layer 2 — analytical

A1 Reason category — MIXED (MANAGEMENT + COMPENSATION). (unstable) Primary — expertise devalued by a management decision (X1, X2), but the respondent herself isn't sure what tipped the scale in choosing the new place. Quote: "Both, honestly I don't know which one tipped the scale, probably predictability in the end. Though the money is nice too, no denying it." — line 40 ⚑ Additional ambiguity (latent construct): the respondent cannot clearly separate burnout as an independent cause from burnout as a consequence of management decisions — line 54 ("whether one causes the other or they're two separate things"). A candidate for human adjudication.

A2 Preventability — PREVENTABLE. Basis: management knew in advance, three times (November, January, April — see X7), but the system didn't respond until the point when the decision became irreversible. Quote: "three times, November, January, April — I wasn't heard" — line 74

A3 Retention signal Specific condition: not a one-off concession, but a durable reprioritization of the budget with a guarantee it would stay that way going forward (not a one-time restoration of the project, but a systemic change). Quote: "What mattered to me wasn't a one-off concession, but knowing it would stay that way going forward." — line 62 Also important: the respondent confirms that an apology without a change in priorities would no longer have worked (line 64) — the retention window closed after January.

A4 Systemic pattern vs individual case — MIXED, LEANING TOWARD SYSTEMIC PATTERN. (unstable) Colleague Katya is also considering options for similar reasons (uneven workload distribution), the respondent herself isn't sure of a direct causal link to her own departure. Quote: "Katya is definitely considering options, but I don't think it's directly because of me, she's got her own similar reasons, the workload distribution hits everyone the same way." — line 72

A5 Reputational echo risk — NEUTRAL-LOYAL TONE, willing to name the substance under direct questioning. Publicly — the moving story (loyal, no bridge-burning); to a direct personal question from the industry — neutral, no mudslinging, but with a specific formulation of the core conflict over the management approach to priorities. Quote: "I'll be honest but neutral. I won't badmouth anyone, I'll just say we didn't see eye to eye on the management approach to priorities." — line 70

A6 Main insight The employee is leaving because her expertise was devalued by a management decision (reorganizing her position into an administrative-procurement role); the decision became irreversible on May 14th at the reorganization meeting; preventability — PREVENTABLE, since signals reached management three times (November/January/April) with no response. Quote: "the point of no return came on May 14th at that meeting, and the trigger was the reagent budget cut on May 17th" — line 74

Reliability summary

  • All Layer 1 and Layer 2 quotes are verified_exact (23/23), see verify_quotes.py.
  • Unstable Layer 2 cells requiring advisory input (S3) in a real run: A1 (reason category — mixed), A4 (systemic pattern vs individual case), and the latent fork "burnout as cause vs consequence" (see ambiguity note under A1).
  • Respondent self-contradiction recorded (line 52) — not "fixed" by the interpreter, but flagged explicitly.