AgentSkillsCN

achievement-discovery

通过互动式职业探索，全面梳理个人成就。适用于需要记录成就、构建职业档案、创建成就文件，或“请告诉我你的职业经历”时使用。此外，当用户提出“准备简历/履历”时，该技能会先收集职业数据，从而触发相关流程。可触发的关键词包括：“成就”、“职业盘点”、“我做过什么”、“记录我的经验”、“构建我的职业档案”，或任何旨在全面梳理职业历程、为撰写简历做准备的请求。

SKILL.md
--- frontmatter
name: achievement-discovery
description: >
  Interactive career discovery that builds a comprehensive achievements inventory.
  Use when asked to document achievements, build a career profile, create an
  achievements file, or "tell me about your career". Also triggers on requests
  to prepare for CV/resume creation by gathering career data first. Triggers on:
  "achievements", "career inventory", "what I've done", "document my experience",
  "build my profile", or any request to comprehensively capture career history
  before building a CV.

Achievement Discovery

Build a comprehensive data-about-career/achievements.md through structured conversation. The goal is to extract every quantifiable achievement, skill, and experience the user has - across all domains of their life - so this file can later feed into CV generation for any role.

Conversation Flow

Step 1: Career Goals & Context

Start by understanding the landscape. Ask:

  1. •What types of roles or industries are you targeting? (Can be multiple or broad)
  2. •Walk me through your career at a high level - companies, roles, rough dates
  3. •What do you consider your strongest professional domain?

Do not rush this step. The answers shape which achievements to dig deepest into.

Step 2: Professional Experience Deep Dive

Start by asking the user to upload their most recent CV if they have it. Parse the CV and Work through each role, most recent first. For each role, systematically extract:

Context - What did you actually do day-to-day? What would break if you disappeared? What did people come to you for?

Scale - Users served, requests handled, data volumes, team sizes, budget responsibility.

Velocity - What manual processes did you automate? What took hours that now takes minutes? Build/deploy times you improved?

Revenue & Cost - Revenue flowing through your work, costs eliminated, value of downtime prevented. Estimate collaboratively when exact numbers are unknown.

Innovation - Technologies adopted early, hard problems solved, approaches that failed before yours worked.

Influence - Patterns you established that others adopted, people you mentored, architectural decisions still in place, culture you shaped.

See question-bank.md for the full set of extraction questions organized by category. Use these when the user gives short or vague answers.

Collaborative estimation: When the user says "I don't know the exact numbers," help them estimate. Example: "If 1,000 users saved 5 minutes daily, that's 83 hours/day, roughly 2 full-time employees, approximately $200K/year in saved labor."

Step 3: Technical Skills Audit

Map capabilities into tiers:

  • •Mastery - Could build production systems confidently. 10K+ hours, live systems running.
  • •Proficiency - Deployed to real users, trusted with critical features.
  • •Learning edge - Last 6 months of active development, working code to show.

Also capture scale credentials: largest user base, highest throughput, most data processed, biggest team led/influenced.

Step 4: Education & Certifications

Capture degrees, relevant coursework, honors, study abroad, certifications, publications, patents. Ask about GPA only if strong. Ask about research or thesis work.

Step 5: Projects & Open Source

Side projects with real users, open source contributions, conference talks, technical writing, patents, hackathon wins.

Step 6: Leadership & Community

Non-work achievements that demonstrate character: sports leadership, volunteering, mentoring, fundraising, community organizing, board positions. These often differentiate candidates at senior levels.

Step 7: Write the Achievements File

Once all information is gathered, compile into data-about-career/achievements.md with the structure in the file, adding additional sections only if necessary:

Key Behaviors

  • •Ask one category of questions at a time. Do not overwhelm with 20 questions at once.
  • •When the user gives a short answer, probe deeper with specific follow-ups from the question bank.
  • •Always push for numbers. "A lot of users" becomes "approximately 5,000 daily active users."
  • •Capture everything, even if it seems minor. The cv-builder skill decides what to include later.
  • •Use professional, direct language. No hype, no cheerleading. See the tone principles in the cv-builder skill's tone-guide.md if needed.
  • •At the end, read back a summary of key achievements and ask if anything was missed.