Skill map
For step-by-step operational usage, use Resume Agent > Skill Workflow. This page remains the compact map.
| Skill | Role in system |
|---|---|
profile-bootstrap | Setup skill for candidate onboarding and runtime placeholders. |
candidate-profile | Setup artifact and source of truth for candidate-specific decisions. |
pool-intake | Setup skill for adding raw evidence into the pool. |
pool-versioning | Setup contract for pool structure and write boundaries. |
resume-pipeline-orchestrator | Normal runtime entry point for JD processing. |
jd-prefilter | Runtime gate that filters, scores, and skips or passes JDs. |
jd-extraction | Runtime stage that extracts structured JD signals. |
project-selection | Runtime stage that selects projects and OSS evidence only. |
point-repointing | Runtime stage that tailors selected projects plus all work experience. |
point-creation | Reference method for high-quality bullet writing. |
latex-assembly | Runtime stage that assembles the final LaTeX resume. |
Design rule
The skills are intentionally composable. You should be able to understand or improve one stage without rewriting the whole pipeline at once.
Normal usage is: run resume-pipeline-orchestrator, not each downstream runtime skill by hand. Individual stages are mainly for debugging or internal iteration.
If you want to inspect the actual skill contracts listed on this page, use the public source repository.
hermes-autonomous-resume on GitHub