Orchestrate a multi-persona panel to synthesize adversarial second opinions on decisions.
npx skills add https://github.com/dividedby/skills --skill council-dividedby
Please help me install this Agent Skill. Skill: council Source: https://github.com/dividedby/skills/tree/main/skills/engineering/council Command: npx skills add https://github.com/dividedby/skills --skill council-dividedby
This skill solves the problem of cognitive bias and blind spots in critical decision-making by orchestrating a diverse, adversarial panel of AI personas to stress-test your proposals.
Invoke the council skill to evaluate the proposed architectural change for the billing module by providing the task description as an argument.
High-intent search queries and answers about installing and using this skill.
To get an adversarial code review for architectural trade-offs, you can use an orchestration tool that assembles a multi-persona panel to stress-test proposals and identify cognitive bias. This process applies diverse evaluative lenses to your design to surface hidden risks.
Reducing cognitive bias in high-stakes architectural decisions requires an anonymized peer-ranking mechanism where multiple diverse personas evaluate your proposal. This structured synthesis forces a mandatory dissent block so you can see rejected arguments and blind spots.
Yes, you can use a multi-persona panel to review your PR by running parallel evaluations with personas like Falsifier, Minimalist, and Pragmatist. The final chair synthesis provides a structured recommendation that includes mandatory dissenting opinions.
Yes, the council skill requires the built-in Workflow tool to manage parallel evaluation, anonymized peer-ranking, and final chair synthesis. This dependency is necessary to orchestrate the multi-persona panel and generate structured recommendations.
You need an adversarial panel for strategic planning when finalizing high-impact product directions or architectural changes where cognitive bias poses a significant risk. It synthesizes diverse perspectives to ensure your strategy is robust against various evaluative lenses.
A blind multi-persona panel synthesizes architectural feedback by running parallel evaluations from distinct personas like Pragmatist and Falsifier, followed by anonymized peer-ranking. The chair then produces a final recommendation with a mandatory dissent block.
Ankh-StudioStructures debates among multiple personas to resolve architectural design choices.
meaganewallerReview technical designs for risks, assumptions, and scalability concerns.
mbruhlerCoordinate multi-agent workflows with sequential, parallel, and conditional execution.
Cogni-AI-OUReview pull requests for logic, architecture, performance, and security issues.
