construct-defense

SkillDev tools

Lets your agent build a strong, evidence-backed case for a position while honestly flagging its weak points.

Available today. Use it from your connected AI after setup.

Add ahel to your AI once: Claude, ChatGPT, Cursor, Claude Code or Codex. Then ask it to use this.

Then ask your AI: use the construct-defense skill

About this skill

Construct the strongest defensible case while acknowledging valid weaknesses.

What this skill tells your AI

The instructions your AI receives, as published by yogsoth-ai/de-anthropocentric-research-engine in skills/construct-defense/SKILL.md and read by ahel’s review.

Purpose

Construct the strongest defensible case while acknowledging valid weaknesses.

Input contract

required: [position, evidence_set, objection_set]
optional: [defense_criteria, audience, weakness_policy]
constraints: [support must be evidence-linked and weaknesses cannot be erased]

Procedure

  1. State the position and burden of proof.
  2. Select and order the strongest supporting arguments.
  3. Address objections and reframe only where evidence supports the reframe.
  4. Mark residual weaknesses and scope limits.

Output contract

produces: [defense_case, supporting_arguments, objection_responses, residual_weaknesses]
delta_fields: [findings, decisions, uncertainties]

Quality gates

  • At least 3 supporting arguments are supplied when evidence permits.
  • At least 1 perceived weakness is tested for legitimate reframing.
  • Every claim has an evidence or uncertainty link.

Parameterization

Caller supplies position schema, evidence rubric, objection taxonomy, argument-count rule, and weakness policy.

Failure and counterexamples

Reject hedged advocacy, unsupported reframes, or defenses that omit valid objections.

Provenance map

  • concept: stress-test/debate-defender
  • concept: convergence/advocate-construction

Signals

GitHub stars
501
Forks
41
Last commit
Sep 2026
Advanced
Catalog kind
skill
Key
construct-defense
Source
github.com/yogsoth-ai/de-anthropocentric-research-engine