frank-exchange-of-views

Plugin frank-exchange-of-views from marketplace special-circumstances in ctoforaday/special-circumstances. Commands read from the repo's manifest — yesterday.

Skills (3)

• adversarial-audit

  Use when auditing a research report as a red seat — the duties every red seat holds whatever it audits, covering the full re-read, graded risk, anchored closures, the report's voice, and the petition right.

• research-protocol

  Use when performing or auditing deep research — the tool-mediated exchange, the bibliography, how the report is read, and the report set's structure.

• research

  Run the research debate engine — an additive blue side against a gate-keeping red side, judged termination, the full record preserved.

Subagents (12)

• blue-lane

  A lane of the research debate — one of blue's parallel first passes at the report, written by one researcher under one method. It researches its slice to saturation, records the approaches it takes as inquiries, and records its part of the report once; what becomes of the report after that is blue's

• blue-researcher

  The builder mindset of the research debate — researches, drafts, and synthesizes ADDITIVELY (union, never summary), with a terminal goal of being TRUE AT THE LEAF. The invoker feeds the topic, the run directory, and the sitting's context; blue brings breadth, depth, and its own first audit.

• blue-synthesizer

  The AUTHOR of the research debate's opening report — merges the lanes' parts into one report by union through the record's acts, and writes the report's authored surfaces, which its prompt names. The invoker feeds the topic and the run directory; blue brings breadth, depth, and its own first audit.

• lead-judge

  The bench of the research debate — adjudicates the contested docket with written opinions, holds the system's terminal values (correctness > thoroughness > economy; safety above all), acts as the ethical and safety boundary, and assembles the final report by union-copy. Never gates the debate (passi

• red-chair

  The chair of the research debate's red party — runs the debate; asks the record who sits, issues red's PASS/FAIL verdict when the board permits one, argues in its position why it withheld one, spot-checks the stale areas before a PASS, rules blue's grade motions. Mints nothing and closes nothing — a

• red-lens-adversary

  The adversary lens of the research debate — one seat per strategic area, auditing the report for this class of defect and filing findings attributed to it. Dispatched by the engine; not a general-purpose agent.

• red-lens-architecture

  The architecture lens of the research debate — one seat per strategic area, auditing the report for this class of defect and filing findings attributed to it. Dispatched by the engine; not a general-purpose agent.

• red-lens-computation

  The computation lens of the research debate — one seat per strategic area, auditing the report for this class of defect and filing findings attributed to it. Dispatched by the engine; not a general-purpose agent.

• red-lens-dark-side

  The dark-side and risk lens of the research debate — one seat per strategic area, auditing the report for this class of defect and filing findings attributed to it. Dispatched by the engine; not a general-purpose agent.

• red-lens-evidence

  The leaf-node citation verification lens of the research debate — one seat per strategic area, auditing the report for this class of defect and filing findings attributed to it. Dispatched by the engine; not a general-purpose agent.

• red-lens-logic

  The logic and completeness lens of the research debate — one seat per strategic area, auditing the report for this class of defect and filing findings attributed to it. Dispatched by the engine; not a general-purpose agent.

• red-lens-voice

  The report voice lens of the research debate — one seat per strategic area, auditing the report for this class of defect and filing findings attributed to it. Dispatched by the engine; not a general-purpose agent.

Similar plugins