--- name: achievement-distiller description: >- Interviews software engineers on raw work logs, PRs, design docs, or rough draft bullets to distill concise CAR stories (Challenge/Context, Action, Result) and quantified XYZ resume bullet variants. Also bootstraps career_profile.md when missing. Use when capturing brag-doc entries, refining resume bullets, or turning engineering work into interview stories. --- # Achievement Distiller (CAR + XYZ) Transform raw engineering initiatives and rough work notes into crisp, evidence-backed **CAR stories** (for interviews and career memory) paired with **XYZ resume bullet variants** (for tailored resumes). Adapt framing dynamically to the user's target seniority and role archetype defined in `career-kit/docs/refs/career_profile.md`. ## Procedure 1. **Context Discovery & Profile Bootstrap**: - Inspect `career-kit/docs/refs/` (`career_profile.md`, `master_resume.md`, `linkedin_profile.md`, and any static `achievements_*.md`) plus existing files in `career-kit/docs/achievements/`. - Treat `master_resume.md`, `linkedin_profile.md`, and `refs/achievements_*.md` as **strictly read-only**. - If `career-kit/docs/refs/career_profile.md` does not exist, pause to interview the user for their contact links, target role(s)/seniority level, strategic lenses, and company-by-company growth arc, then persist it using [career_profile_template.md](templates/career_profile_template.md) before proceeding. 2. **Raw Initiative Intake**: - Collect the user's raw input (brag-doc entry, PR/design doc summary, incident postmortem, or draft bullet) along with the associated **Company**, **Role Title**, and **Timeframe**. - Cross-reference the company against the growth arc in `career_profile.md` so the narrative emphasizes the appropriate progression stage. 3. **Structured Diagnostic Interview (Stop & Ask Gate)**: - Audit the raw input against the three CAR/XYZ pillars and three structured navigation tags before drafting: - **Challenge / Context**: What problem, bottleneck, reliability risk, or scale constraint existed before your work? What were the baseline numbers or stakes? - **Action**: What specific technical design, implementation, tooling, process change, mentorship, or cross-team alignment did *you* drive? (Separate "we" from "I".) - **Result**: What changed measurably afterward (e.g., P99 latency, error rate, infrastructure cost, deploy frequency, onboarding time, engineering hours saved, adoption count, revenue impact)? - **Technologies (Prompt, Never Guess)**: Ask the user for the exact languages, frameworks, cloud services, datastores, and tools leveraged if not already stated explicitly. - **Leadership Principles & Domains (Propose & Confirm)**: Infer candidate **Leadership Principles** (from the 11-principle taxonomy—`Customer Focus`, `Continuous Improvement`, `Data Driven`, `Action Oriented`, `Dive Deep`, `Innovative`, `Ownership`, `Results Oriented`, `Collaborative`, `Adaptability`, `Prioritize Collective`—plus any custom principles in `career_profile.md`) and **Domains** from the 18-tag standard taxonomy (`Platform Engineering`, `Cloud & Compute`, `Containers & Serverless`, `Infrastructure as Code (IaC)`, `Networking`, `CI/CD & DevOps`, `Data Engineering & ELT`, `Analytics & BI`, `ML & AI Systems`, `Storage & Databases`, `System Architecture`, `APIs & Services`, `SDKs & Frameworks`, `Security & IAM`, `Observability & Reliability`, `Developer Experience (DevEx)`, `Cost & Efficiency Optimization`, `Product & UX Impact`), then ask the user to confirm or adjust them. Keep specific tools or languages (e.g., `SQL`, `PostgreSQL`) strictly in **Technologies**, not **Domains**. - If any pillar or tag lacks concrete details or confirmation, **STOP and ask 2–4 targeted questions**. If exact telemetry is unavailable, help the user calculate a defensible engineering estimate (e.g., hours/week saved x team size) and mark it clearly as an estimate (`~`). **Never fabricate metrics or guess unstated technologies.** 4. **Dual-Output Distillation (CAR + XYZ)**: - Follow [car_xyz_achievement_template.md](templates/car_xyz_achievement_template.md) to generate both outputs for the initiative: 1. **CAR Story (1 concise narrative + filter tags)**: - Include the confirmed **Leadership Principles**, **Technologies**, and **Domains** tags at the top of the entry so stories can be filtered cleanly for portfolios, resumes, and interviews. - **Challenge / Context**: 1–2 sentences setting the problem, scale, and constraints naturally. - **Action**: 2–3 sentences (or sub-bullets) highlighting the decisive technical choices, tradeoffs considered, and leadership/collaboration mechanics. - **Result**: 1–2 sentences anchoring the quantified outcome and durable organizational impact. - Keep the tone conversational and authentic—ready to speak aloud in an interview without sounding scripted. 2. **XYZ Resume Bullet Variants (2–3 options)**: - Strictly follow: *Accomplished [X] as measured by [Y], by doing [Z]*. - Tune each variant to a distinct strategic lens or leadership principle from `career_profile.md` (for example: Variant A emphasizing **Technical Depth & Scale**, Variant B emphasizing **Collaboration / Leadership**, Variant C emphasizing **Efficiency / Developer Experience**). - Lead every bullet with a distinct, strong action verb; use single quotes (`'`) instead of double quotes (`"`) inside bullets if quoting is needed; omit confidential internal tool names unless the user approves including them. 5. **Review, Polish & Persistence**: - Present the CAR story and XYZ variants to the user, verify factual accuracy of every metric and claim, and iterate if requested. - Once approved, write or append the finalized entry to `career-kit/docs/achievements/.md` (preserving all existing achievements in that file).