Evaluate convergence gaps and recommend deterministic next work for project targets.
npx skills add https://github.com/marcelocantos/skills --skill cv-marcelocantos
Please help me install this Agent Skill. Skill: cv Source: https://github.com/marcelocantos/skills/tree/main/skills/cv Command: npx skills add https://github.com/marcelocantos/skills --skill cv-marcelocantos
Identifies convergence gaps and produces a deterministic, prioritized next action across a project's active targets so engineers do not have to manually triage frontier targets, unreleased fixes, and standing invariants.
Run the cv skill at the project root to compute convergence, display the full structured recommendation, and proceed with the Next action if you want the agent to begin work.
High-intent search queries and answers about installing and using this skill.
To determine the next work for active project targets, you evaluate convergence gaps to produce a deterministic, prioritized next action. This surfaces the top frontier target and can automatically begin the first acceptance-driven change.
Convergence gap evaluation is the process of identifying unreleased fixes and standing invariants across active project targets. It produces structured summaries to help engineers triage frontier targets without manual effort.
Yes, you can run a global portfolio mode to rank repositories and highlight cross-repo blockers. This mode evaluates convergence gaps across multiple repositories and optionally applies momentum weighting from recent activity.
Yes, target prioritization requires the bullseye_convergence tool to be available in the environment. It also optionally uses mnemo_recent_activity for momentum data and requires adherence to release gating wrappers.
When unreleased fixes are detected, the deterministic next action invokes gated skills such as the release process rather than bypassing release gates. This ensures fixes are properly released.
The best way to triage frontier targets without manual effort is to run an automated evaluation at the repository root. This computes convergence, displays a structured recommendation, and proceeds with the next action.
scooter-lacroixSolve optimization problems with gradient descent, momentum, and Newton's methods.
ZardLi1115Decompose requests into Kanban task graphs routed to specialist profiles.
linenoizeDeploy applications to Vercel, Netlify, Fly.io, Docker, or VPS with verification and rollback.
gugug168Retrieve and refine relevant codebase files through iterative dispatch-evaluate-refine loops.
