Total Skills
13
Skills published by ViperJuice with real stars/downloads and source-aware metadata.
Total Skills
13
Total Stars
0
Total Downloads
0
Comparison chart based on real stars and downloads signals from source data.
skill-editor
0
skill-improvement-planner
0
phase-roadmap-builder
0
plan-phase
0
page-load-monitor
0
smart-screenshot
0
execute-phase
0
batch-verify
0
Apply an improvement plan produced by skill-improvement-planner to the relevant SKILL.md files. Archives the reflection files consumed by the plan so they aren't re-aggregated next cycle. Use when the user says 'apply the skill improvement plan', 'update the skills with the aggregated feedback', 'run the skill editor', or follows a skill-improvement-planner run. Ingests the plan at ~/.claude/skills/skill-improvement-planner/plans/plan-v<N>-<ISO>.md (or the path in the predecessor handoff), applies each recommendation via a frontier-tier agent, mirrors edits across dotfiles and team-skills repos where applicable, and commits each edited repo.
Aggregate reflection files from the planning-chain skills (phase-roadmap-builder, plan-phase, execute-phase) and produce an improvement plan that a downstream skill-editor will ingest. Use when the user wants to review accumulated skill feedback, aggregate reflections, plan skill improvements, or asks 'what changes should I make to my skills' — phrases like 'review my reflections', 'aggregate skill feedback', 'plan improvements to the pipeline skills', 'let's see what the reflections say', or after several phase executions have run and the reflections have accumulated. Does NOT edit skills itself — produces a plan for a separate skill-editor skill to apply.
Produces or extends a multi-phase roadmap spec (`specs/phase-plans-v<N>.md`) that `/plan-phase` can ingest. Use when the user wants to turn the current conversation's plan, an architecture discussion, or a pointed-at markdown spec into a phased roadmap — phrases like 'build a phase roadmap', 'turn this into a phased plan', 'create a multi-phase spec', 'add a phase to the roadmap', 'extend the phase plan', or after a long planning conversation when the user says 'ok let's formalize this'. Also triggers when the user wants to APPEND phases to an existing roadmap without editing prior phases. Designs for maximum parallelism: fewer phases, more parallel lanes per phase, tight early interface freezes.
Architecture-first planning for a spec phase. Produces an interface-freeze + swim-lane document for parallel execution. Use in plan mode. Supports --consensus for multi-agent architectural consensus across named Plan teammates.
Intelligent browser page load failure handling. Use when browser screenshot or navigation operations fail, especially with 'Frame showing error page' or timeout errors. Provides escalating diagnostic steps (console, network, alternatives) instead of blind retry loops.
Guidelines for efficient browser screenshot usage. Use when performing browser automation tasks involving screenshots, especially in UI testing sessions. Provides strategies to avoid redundant screenshot captures, intelligent wait strategies for page loads, and alternatives when screenshots fail repeatedly.
Executes a plan-phase plan doc to completion. Spawns one worktree-isolated teammate per swim lane via TeamCreate, respects the lane DAG and interface freeze gates, auto-merges lanes when their verify passes, and retries once on failure before halting. Assigns each teammate a model + thinking level matched to the lane's complexity. Run after /plan-phase has produced plans/phase-plan-<VERSION>-<PHASE_ALIAS>.md.
Workflow for batching edit verification instead of checking after each individual change. Use when making multiple related file edits (3+ files) in a coding session. Groups syntax checks, type checking, and test runs into a single verification phase after all edits are complete, instead of scattering verification calls throughout.
Diagnostic guide for Bash command failures. Use when a Bash command returns a non-zero exit code. Provides error categorization by exit code, root cause analysis procedures, and specific fix suggestions. Helps avoid blind retry loops by diagnosing the error before attempting another command. Trigger on any Bash failure, especially exit codes 1, 2, 127, and 255.
Guidelines for avoiding redundant file reads during a session. Use throughout any coding session. Instructs Claude to track which files have been read and avoid re-reading unchanged files.
Checklist for safe file editing operations. Use before any Edit or Write tool call. Enforces read-before-edit pattern, detects potential conflicts from external modifications, and prevents common edit failures.
Guide for preparing rich context when spawning subagent tasks. Use before creating any Task tool call, especially for Explore or general-purpose subagents. Provides templates for task prompts that include necessary file paths, architecture context, and scope boundaries.
When to trigger this skill - be specific about the scenarios where Claude should use it