Orchestrate multi-persona adversarial reviews of engineering artifacts and decisions.
npx skills add https://github.com/Zhachory1/agent-fleet --skill council-zhachory1
Please help me install this Agent Skill. Skill: council Source: https://github.com/Zhachory1/agent-fleet/tree/main/skills/council Command: npx skills add https://github.com/Zhachory1/agent-fleet --skill council-zhachory1
This Skill solves the problem of cognitive bias and oversight in high-stakes engineering decisions by convening a diverse, adversarial council of specialist personas to critique your work before you commit to a path.
Invoke the council skill to perform a deep-dive review of the current design document by analyzing it with a team of specialist personas.
High-intent search queries and answers about installing and using this skill.
An adversarial architectural design review convenes a diverse council of specialist personas to critique your work before committing. It orchestrates a bounded, multi-round debate to catch flaws a single reviewer would miss.
Evaluating a build-vs-buy decision without cognitive bias requires convening an adversarial council to critique the strategic choice. This process produces a decision-grade report featuring ranked issues, named dissents, and a false-consensus flag.
Multi-persona debate for engineering reviews works by orchestrating a bounded, multi-round discussion between 3-6 independent personas. It applies iterative reflection rounds to mitigate false consensus and surface scope-creep risks.
Yes, you can use persona-based reasoning to critique a high-stakes pull request. The process evaluates the PR by applying specialist personas to ensure you have considered all failure modes and scope-creep risks.
Yes, adversarial code review requires a structured environment with persistent room capture. This environment supports persona-based reasoning and iterative reflection rounds necessary to produce a decision-grade report.
You should not use a multi-persona council for routine engineering decisions that do not warrant a multi-round debate. This approach is reserved for high-stakes artifacts, major architectural changes, and strategic assessments to prevent oversight.
jsagirDetect gaps, contradictions, and convergence in structured data room entries.
404-error-notfoundAnalyze code changes against CLAUDE.md guidelines and output a structured remediation report.
inho-teamStructure multi-round, multi-agent debates to surface tradeoffs and decision rationale.
open-horizon-labsSurface contrary evidence and assumptions during major decisions.
