Candidacy Checker Eval

candidacy-checker-eval

Evaluate job application readiness before submitting.

When to use this skill

• A candidate wants to know if they are ready to apply for a specific role.
• An AI assistant is about to help write a cover letter or resume and needs

to audit what materials actually exist vs. what the candidate claims.

• A candidate is deciding between applying now vs. spending more time on

preparation.

• A hiring coach wants a structured gap report for a client.

When NOT to use this skill

• After the application is submitted (the framework is pre-submission).
• For evaluating interview performance (different rubric needed).
• For comparing two candidates against each other (single-candidate scope).

Mechanism

See prompts/ for the four prompt files that drive the workflow:

1. 00_kickoff.md -- set frame and intent
2. 01_profile_discovery_loop.md -- Socratic extraction with Trinity Dialectic halt gates
3. 02_rubric_runner.md -- flat boolean checklist execution
4. 03_remediation_planner.md -- sequence FAIL/UNKNOWN rows into an action plan

Output

• Populated templates/role_profile_template.md
• Populated templates/candidate_profile_template.md
• Populated templates/checklist_template.md with per-row verdicts
• Populated templates/action_plan_template.md with sequenced fixes

Citations

• agentskills.io spec
• AGENTS.md spec (Linux Foundation Agentic AI Foundation)
• Adaptive Precise Boolean Rubrics, arXiv:2503.23339
• Checklists Are Better Than Reward Models, arXiv:2507.18624
• Three-lens dialectical evaluation: Logos/Pathos/Ethos gating (Aristotle's Rhetoric, Books I-II)
• Zero-assumption mandate and halt-on-doubt confidence loop (agentic system design pattern)

Source & license

This open-source skill is cataloged on AgentStack and links to its original source — we do not rehost the code.

• Author: 0SxD
• Source: 0SxD/candidacy-checker-eval
• License: MIT

Install and usage instructions live in the source repository linked above.