vellum-feature-flag-rollout
Guide Vellum Assistant feature flag changes and rollout hygiene. Use when adding, editing, reviewing, or documenting assistant feature flags, rollout-gated behavior, or platform flag follow-up work.
Browse reusable Agent Skills, each with a clear purpose and practical guidance.
Guide Vellum Assistant feature flag changes and rollout hygiene. Use when adding, editing, reviewing, or documenting assistant feature flags, rollout-gated behavior, or platform flag follow-up work.
Author and review first-party Vellum skills in the repository. Use when editing files under skills/, assistant bundled skills, skill manifests, tool definitions, SKILL.md files, or first-party skill portability and isolation behavior.
Discover bundled skills and search/install community skills from the skills.sh registry
Complete voice configuration in chat - PTT key, microphone permissions, ElevenLabs TTS, and troubleshooting
Guide instance administrators through AI provider setup, model configuration, MCP server management, and instance settings.
Analyzes Guidance evaluation results to identify patterns and causes of low pass rates. Use when investigating low guided pass rates, checking trajectories, or debugging eval grader failures.
Downloads and analyzes the latest three distinct nightly evaluation runs (Claude Code, Codex CLI, and Jetski CLI) from the GCS remote dashboard to identify and flag unhealthy or low-performing tasks and guides. Use this skill whenever you need to run a bulk investigation on remote nightly runs, track agent health, or identify over-prescribed/brittle guides.
Add or update Curvine cv-* project skills under .agents/skills/, following naming and layout conventions, and sync the AGENTS.md registry. Use when creating a new cv skill, updating an existing cv skill structure, or registering skills after changes.
Update CLAUDE.md and skills when introducing new patterns or conventions. Triggers: "update claude", "update skills", "new pattern", "add convention".
Reflect on completed research tasks to improve future performance. Use when: a research task has just been completed and the agent should evaluate its own process, store lessons learned, or retrieve past reflections before starting new work. NOT for: active research execution or data analysis.