isizono
Community@isizono
isizono builds CALM/cc-memory, a persistent memory and decision-tracking system for Claude coding sessions via a local MCP server.
Agent Skills by isizono
Showing 35 vetted skills indexed across 2 GitHub repositories.
task-execute
Executes implementation plans via git worktrees, subagents, reviews, and PR creation.
task-plan
Creates implementation plans with PR splitting and edge-case coverage for work tasks.
man
Explains CALM usage, record model, and skill index to users and Claude.
tag-notes
View and update persistent notes attached to tags in the CALM memory database.
postmortem
Reviews completed activities step by step to extract and persist lessons learned.
coding-project-setup
Set up a persistent knowledge frame for a coding project in CALM.
tag-cleanup
Analyze tag co-occurrence and execute user-approved tag consolidation actions.
rule-placement
Determines the correct context-delivery channel for generalized agent behavior rules.
check-in
Retrieves aggregated context and progress for an existing activity via the check_in tool.
board
Posts asynchronous questions, announcements, and discussion topics to shared bulletin board topics.
peer-nudge
Governs synchronous SendMessage outreach to peer sessions listed in the session ledger.
ask-answer
Lists open asks and records human answers via answer_ask in a linear flow.
ask-compose
Structures question and context fields for asynchronous decision requests before calling add_ask.
recording
Records discussion logs and session artifacts into a persistent memory database.
digest
Generates cross-period digests of activities, decisions, and materials grouped by domain.
memory-import
Imports exported memory bundles into a CALM database with conflict adjudication.
ask-distill
Aggregates repeated ask rulings into a meta-ask proposing self-adjudication precedent.
setup-anchor
Define or update verification anchors for recorded decisions, logs, and materials through user dialogue.
restart
Restarts the CALM local MCP server process to apply plugin code updates.
project-setup
Set up knowledge frames for new project domains in the CALM memory database.
decision-record
Records user-approved decisions and unresolved discussion points via add_decisions in CALM.
db-recovery
Restores CALM SQLite database from snapshots after detecting abnormal row count decreases.
activity-pause
Pauses in-progress activities with resumption notes and status transitions in CALM.
audit
Reconstructs decision history from logs and primary sources to analyze context gaps and pin corrective knowledge.
Frequently Asked Questions About isizono
FAQPage SchemaWhat tasks can I automate using isizono's CALM skills?▼
You can persist session knowledge as topics, decisions, logs, and materials; start, pause, check in on, and finish activities; generate ADRs and meeting notes with scribe; run postmortems; plan and execute PR-split implementation tasks; and export or import memory bundles between cc-memory instances.
Who is the target user for isizono's skill registry?▼
Engineers running Claude Code sessions who need durable cross-session memory. Personas include developers managing long-running coding projects, reviewers auditing past decisions, and operators coordinating asynchronous asks and message boards between multiple concurrent Claude sessions.
How does a typical CALM runtime session work?▼
Skills trigger on user intent or autonomously: activity-start opens work, recording and decision-record persist context as it happens, check-in resumes prior activities, and sync-memory consolidates transcripts at session end. The restart skill reloads the local MCP server after plugin updates.
What are the prerequisites and dependencies for CALM?▼
CALM requires a running local MCP server (restartable via the restart skill, with optional --restart-embedding for the embedding server), a database whose integrity is monitored by SessionStart hooks, and git worktree support for the task-execute implementation flow.
How does CALM handle data loss or stale memories?▼
The db-recovery skill detects abnormal database row-count drops and guides snapshot-based full or partial restoration with re-verification. The forget skill retracts outdated decisions, habits, and tag-notes, while audit re-evaluates contradicted past decisions and re-pins knowledge correctly.