judge-resume

Prompt file imported from Dharmik7601/TailorTex (.claude/commands/judge-resume.md). Fill in {{arguments}} before use. Copyright stays with the author.

Allowed file operations: Read any file listed in the steps below. Write ONLY to prompts/daily_feedback.json. Do not create any new files or write to any other path under any circumstance.

Evaluate the resume output/extras/$ARGUMENTS_Resume.txt against the current job description.

Follow these steps exactly:

1. Read output/extras/$ARGUMENTS_Resume.txt — this is the plain-text extract of the resume, parsed from the LaTeX source by the backend before invoking this command; it contains Technical Skills, Experience, Projects, and Education in clean readable format with no binary artifacts

2. Read output/extras/$ARGUMENTS_jd.txt — this is the job description snapshot written by the backend specifically for this evaluation run; it is isolated per-job and guaranteed to match the JD used to generate this resume

3. Read prompts/prompt_summary.txt — this contains the condensed rule reference used during generation (use as the system_prompt / user_constraints context for the evaluator)

4. Read prompts/evaluator_prompt.txt — this contains the role, scoring categories, output schema, and rule_reference_map

5. Apply the full evaluation rubric from prompts/evaluator_prompt.txt to the resume text from step 1, using:

   • output/$ARGUMENTS_jd.txt content as the job_description input
   • prompts/prompt_summary.txt content as the system_prompt and user_constraints context
   • Set resume_id to {{arguments}}

6. Read prompts/daily_feedback.json if it exists:

   • If it exists and contains a valid JSON array, append the full evaluation result object to that array
   • If the file does not exist, is empty, or does not contain a valid JSON array, create a new array containing only the evaluation result object
   • Write the updated array back to prompts/daily_feedback.json

7. Output ONLY the following JSON object to stdout (nothing else — no markdown, no prose): {"total_score": <number>}