Skip to main content

Skill map

For step-by-step operational usage, use Resume Agent > Skill Workflow. This page remains the compact map.

SkillRole in system
profile-bootstrapSetup skill for candidate onboarding and runtime placeholders.
candidate-profileSetup artifact and source of truth for candidate-specific decisions.
pool-intakeSetup skill for adding raw evidence into the pool.
pool-versioningSetup contract for pool structure and write boundaries.
resume-pipeline-orchestratorNormal runtime entry point for JD processing.
jd-prefilterRuntime gate that filters, scores, and skips or passes JDs.
jd-extractionRuntime stage that extracts structured JD signals.
project-selectionRuntime stage that selects projects and OSS evidence only.
point-repointingRuntime stage that tailors selected projects plus all work experience.
point-creationReference method for high-quality bullet writing.
latex-assemblyRuntime 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