Appends terse, agent-readable session summaries to a project's living.md file.
npx skills add https://github.com/jpmoya/claude-agents --skill living-jpmoya
Please help me install this Agent Skill. Skill: living Source: https://github.com/jpmoya/claude-agents/tree/main/skills/living Command: npx skills add https://github.com/jpmoya/claude-agents --skill living-jpmoya
High-intent search queries and answers about installing and using this skill.
Say "update living" or "update living.md" and the skill appends a dated entry at the top of the file. It gathers context from recent git commits and the conversation, then writes SHIP, DEC, FIND, OPEN, FILES, and LESSON lines.
Each entry starts with a date and kebab-case slug, followed by tagged one-line facts: SHIP for delivered work, DEC for decisions with reasons, FIND for discoveries, OPEN for unresolved items, FILES for changed paths, and optional LESSON for reusable insights.
Yes, it works for any project by locating living.md at the git repo root via git rev-parse. If the file does not exist, it creates one with a standard header before appending the entry.
The file is written for future agents, not humans, so it optimizes for facts per token. Bold, italics, and filler phrases add markup cost without information, while identifiers like IDs and figures are always preserved verbatim.
No, the skill does not commit living.md after writing. The user commits it manually or folds it into the next pull request, keeping the log under explicit human control.
YanCheng-goSummarize completed session work and update MEMORY.md with progress and blockers.
virtengineLoad and parse STATE.md and PROJECT.md to restore project context.
Bliz243Analyze CLAUDE conversation history to recommend CLAUDE.md file updates.
MDSIXONEEnforces commit message and pull request conventions for GitHub workflows.
