Results for “cdeust”

The harness cockpit, a viewer: the phase every agent is in (PHASES panel, and main and the running agents in the band above the prompt), the advisor calls, what the insights learned (per-phase evidence, session outcomes from /insights, goals, the calibration and its proposals, with buttons to run /insights, apply, or open a calibration goal), Cortex memory health and grooming, the context...

Two rules no settings hook enforces, refused at the tool call: wiki pages and their generated ADR mirrors are written only through the Cortex wiki tool, and every git worktree lives at /.claude/worktrees/ / (one created there is registered with disk_hygiene.py).

/wiki : compile a Cortex wiki page (Markdown with tikz blocks) to PDF through pandoc and xelatex, and open it.

The owner's interface to the plugin fleet: /fleet shows each owned plugin's installed and offered version, the repository's open PRs with their CI state and its open issues (defects first, feature requests set apart), and the session's harness signals. The mod only reads (gh pr list, gh issue list, git remote): every outward action starts with a button the owner presses (Review a PR, Take a...

A graphical pane of the harness at work: every agent of the turn as a swimlane coloured by phase, bar height its effort, with advisor calls and failed tool calls marked; the effort each phase is calibrated to; and each phase's learning curve from the insights ledger, with the calibration's moves. Drawn in pixels where the terminal has the kitty graphics protocol (kitty, Ghostty, outside tmux)...

Automatic effort and model, and a quota that limits how many agents run, not how well one runs. Every request runs at the effort its grade calls for (zetetic-genius) and the main tool loop steps down the ladder. A subagent runs on the model its Agent call names, else the one its agent definition pins (read from the definition file, since agent.spawn carries only the call's), an inheriting type...

Grades every request before its turn starts (task class → effort rung, in one low-effort completion) and routes it to the genius reasoning patterns and problem-shaped skills whose trigger it matches, from the installed zetetic-team-subagents index; under enforce, the matched pattern's workflow rides beside the prompt. An analysis or critical request also carries one line asking for the advisor...

The learning loop: every main turn is kept in a ledger across sessions (phase, effort, steps, failed tool calls, agents, advisor calls), joined with the outcomes observed at tool.call (commits kept or reverted, test exits, PR verdicts and merges, no transcript) and the ones /insights writes, and the project's goals; probe sessions are counted and kept out; a phase whose evidence says so gets...