bdjobs-job-apply（bdjobs 工作申请）

BDJobs Job Apply

Use this skill to help a user manage BDJobs job searching and applying without needing to know APIs or code.

What this skill does

• Save resume text/LaTeX into data/resume.md
• Save credentials and preferences into data/userDetails.json
• Log in and save auth data into data/loggedInData.json
• Refresh applied jobs into data/appliedJobIds.json
• Keep a not-liked list in data/notLikedJobIds.json
• Search jobs and return raw job details for AI matching
• Auto-apply to matched jobs when asked
• Undo/cancel an applied job when asked
• Update expected salary for an already applied job

Onboarding flow

When the user first sets up this skill, ask for these values in plain language:

1. BDJobs username
2. BDJobs password
3. Resume text or LaTeX
4. everyDayApplyCount (default: 2)
5. everyDayCronTime (default: 12:00 PM)
6. isFresher (yes/no)
7. Experience in years
8. Preferred jobLocation (optional, leave blank for all)

Save them in:

• data/userDetails.json
• data/resume.md

If the user has extra search filters or matching rules, store them in data/preferences.json.

Login flow

Before search/apply work, always ensure data/loggedInData.json exists.

If missing:

1. run ChecKUsername
2. take guidId
3. run Login
4. save token, refreshToken, encryptId, decodeId, and guidId to data/loggedInData.json

If login returns 401 at any point:

1. run ChecKUsername again
2. run Login again
3. overwrite data/loggedInData.json

Search flow

When the user asks for fresh jobs:

1. Clear data/suggestedJobs.json
2. Run scripts/bdjobs-refresh-applied.js to update data/appliedJobIds.json from GetApplyPositionInfoV1
3. Run scripts/bdjobs-fetch-jobs.js with --keyword, --isFresher, --postedWithin, --pg, and --jobLocation to fetch raw job lists from GetJobSearch
4. Run scripts/bdjobs-filter-jobs.js to exclude already-applied and not-liked job IDs before Job-Details
5. For selected job IDs, run scripts/bdjobs-job-details.js to fetch full job details
6. Let AI compare the raw job details with data/resume.md and data/preferences.json
7. Run scripts/bdjobs-rank-jobs.js to save the top 5 matched jobs into data/suggestedJobs.json
8. Show the contents of data/suggestedJobs.json to the user

Recommended match threshold:

• show apply suggestions only if match is above 40%
• if the user asks for a looser list, show 30%+

AI matching details:

• Compare the raw job details with data/resume.md and data/preferences.json
• Judge fit from skills, education, age, experience, responsibilities, location, and preferences
• Output a suggested job list with match percentages

Auto-apply flow

When the user asks to auto-apply:

1. Ensure login exists
2. Refresh applied jobs first
3. Search fresh jobs
4. Let AI score the raw job details against the resume and preferences
5. For each approved job:

- call JobApply - read MinimumSalary - use it as expectedSalary in JobApplyPost - call JobApplyPost - if successful, always show matchingScore - if successful, add the job ID to data/appliedJobIds.json - save result in data/lastApplyResult.json

1. Send Telegram notification only after success

Undo/cancel flow

When the user asks to undo an application:

1. Ensure login exists
2. Call UndoJobApply with job ID and FormValue
3. If successful, remove the job ID from data/appliedJobIds.json
4. Save result in data/lastUndoResult.json

Expected salary update flow

When the user asks to update salary for an applied job:

1. Ensure login exists
2. Call UpdateExpectedSalary
3. Save result in data/lastSalaryUpdateResult.json

Files used by this skill

Important files live inside this skill folder:

• data/resume.md
• data/userDetails.json
• data/loggedInData.json
• data/appliedJobIds.json
• data/notLikedJobIds.json
• data/preferences.json
• data/suggestedJobs.json
• data/lastApplyResult.json
• data/lastUndoResult.json
• data/lastSalaryUpdateResult.json

Script entry points

Run scripts from the skill folder, not the workspace root:

• scripts/init-job-profile.js
• scripts/bdjobs-login.js
• scripts/bdjobs-refresh-applied.js
• scripts/bdjobs-fetch-jobs.js

bdjobs-fetch-jobs.js accepts:

• --keyword=...
• --isFresher=true|false
• --postedWithin=... (optional day count, any number)
• --pg=...
• --jobLocation=...
• scripts/bdjobs-filter-jobs.js
• scripts/bdjobs-job-details.js accepts --jobId=... or a positional job id
• scripts/bdjobs-rank-jobs.js
• scripts/bdjobs-apply.js
• scripts/bdjobs-undo.js
• scripts/bdjobs-update-salary.js

Matching rules

Use the resume and these signals:

• Job title
• Company name
• Job description
• Education requirements
• Age requirement
• Experience requirement
• Additional job requirements
• Suggested skills
• Job location
• Job nature
• User preferences from data/preferences.json

Helpful behavior

When presenting jobs to the user:

• keep language simple
• explain only what matters
• include the direct job link
• mention whether it looks worth applying

When the user wants automation, do the work directly with the scripts.