Generates up to 3 memory pin candidates from a session for user approval.
npx skills add https://github.com/Cain-Ish/claude-code-plugin --skill improve-cain-ish
Please help me install this Agent Skill. Skill: improve Source: https://github.com/Cain-Ish/claude-code-plugin/tree/main/skills/improve Command: npx skills add https://github.com/Cain-Ish/claude-code-plugin --skill improve-cain-ish
It helps you capture the most useful decisions, conventions, and blocker outcomes from a session so you do not have to remember them later.
Session-to-pin proposal: Generates up to 3 grounded pin candidates from what happened during the current or most-recent session, covering user feedback, decisions, and blockers. Persona-signal promotion: Detects recurring persona patterns in persona-signals.jsonl and proposes new USER.md rules when they have enough repeated evidence and have not already been captured. User-controlled adoption: Filters candidates for whether you would want them in two weeks, then presents each one for Y/N/edit before applying it via the correct pin/archive destination.
Ask the skill to improve by generating up to 3 candidate pins from the current session and presenting each one for you to accept, reject, or edit.
High-intent search queries and answers about installing and using this skill.
You can externalize session decisions for knowledge management by generating up to 3 grounded memory pins from recent activity, which are then filtered for long-term usefulness and added to USER.md, project facts, or a local wiki.
Externalizing blocker outcomes without autonomous memory involves proposing grounded pin candidates from session signals and optional persona-signals.jsonl, requiring explicit Y/N/edit user approval before applying them via pin_to_user, pin_to_project, or archive_to_wiki tools.
To turn a workflow session into actionable pins, request the Skill to generate up to 3 candidate pins from current session evidence, review each filtered candidate for two-week relevance, and approve, reject, or edit them before they are applied to the correct destination.
Session summarization does not strictly require persona-signals.jsonl; it uses session signals by default to propose pins, but optionally reads persona-signals.jsonl to detect recurring patterns and propose new USER.md rules when enough repeated evidence exists.
Yes, you can promote recurring user feedback patterns into project facts by detecting repeated persona patterns in persona-signals.jsonl, filtering them for long-term usefulness, and pinning the approved candidates to project facts or the local wiki.
You should avoid autonomous session summarization when you need strict user-controlled adoption, as this Skill filters candidates for two-week relevance and presents each one for explicit Y/N/edit approval before using any pin or archive tool.
nsuberiConvert user background and mastery evidence into persistent identity roots.
Cain-IshDetect existing CLAUDE Code and agent files and propose pin updates to USER.md and PROJECT.md.
PedroGiudiceSearch past session observations stored in a local DuckDB memory store.
WalliieeConvert raw memory items into wiki-ready markdown entries with category-specific structure.
