construct-defense
SkillDev toolsLets 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.
No other account needed.
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
- State the position and burden of proof.
- Select and order the strongest supporting arguments.
- Address objections and reframe only where evidence supports the reframe.
- 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