Session Driven Polish — Productivity skill for Claude Code
Use session and promotion evidence to decide which workflow asset deserves the next refinement.
How to install Session Driven Polish
Installs to ~/.claude/skills/ilearn-lab-evoharness-session-driven-polish/SKILL.md
mkdir -p ~/.claude/skills/ilearn-lab-evoharness-session-driven-polish && curl -fsSL https://raw.githubusercontent.com/iLearn-Lab/EvoHarness/HEAD/.claude/skills/session-driven-polish.md -o ~/.claude/skills/ilearn-lab-evoharness-session-driven-polish/SKILL.md Restart Claude Code, or start a new session, for it to be picked up.
What Session Driven Polish does
name: session-driven-polish description: Use session and promotion evidence to decide which workflow asset deserves the next refinement.
Session Driven Polish
- Look at repeated stalls, missing tools, and command drift across archived sessions.
- Distinguish between a one-off bug and a reusable workflow gap.
- Prefer edits that shorten future sessions instead of only improving one transcript.
- If a gap keeps recurring, package the fix as a command, skill, agent, plugin, or MCP asset
Alternatives in Productivity
- Run Discovery — Workflow recipe — go from a vague opportunity to validated insight and a prioritised next step by chaining 4 s 1.3k ★
- Maestro Issue — Intent-driven issue lifecycle management — describe what you want in natural language (报告一个 bug / 列出开放 issue / 530 ★
- Schedule List — List all scheduled tasks with their status and next run times (project) 510 ★
Full documentation available on GitHub
View Source RepositoryRelated Skills
Mine Sessions
Mine every Claude Code session across all projects for the two things that decide a task's cost - how many too
Research Assistant Suite ChatGPT
Self-contained ChatGPT/Codex research workflow bundle for methodology, literature search, and evidence-aware n
Eval
Run eval-driven development workflow. Execute verification loops and evaluate session outcomes against defined
Proof Mode
Delivery harness skill — it runs one task from end to end: resolves its context, routes to the pattern skills,
Log Completion
Log one shipped completion (hive asset, workflow, project, bug fix, etc.) into the user's portfolio completion
Skills Refine
Read the back-and-forth that followed a Skill invocation in the current session and propose edits to fold that
Related Agents
Fix Builder
Fresh session that revises a Blender asset after a critic FAIL. Use only when gauntlet/evidence/ / / contains
Harvest Curator
Draft session learnings from harvest staging into lizard-ready adoption notes. Propose-only; never writes dura
Context Refinement
Updates task context manifest with discoveries from current work session. Reads transcript to understand what