By KavasiMihaly
Curated hooks and tooling for Claude Code: per-session agent action logs, an install-command guard, and a token-aware statusline.
Curated hooks and tooling for everyday Claude Code use. See what your agents are doing. Block what you didn't approve. Track tokens at a glance.

By Mihaly Kavasi — part of the OneDayBI Marketplace.
| Tool | What it does | How it installs |
|---|---|---|
| Agent Logger | JSONL session logs in _Agent Logs/ per repo. Captures prompts, tool calls, subagent activity, and errors. Sensitive paths and prompts redacted. | Auto-wired on /plugin install |
| Block Installs | PreToolUse guard that catches npm/pip/choco/winget/brew and similar install commands and asks for explicit approval first. Approve with a # APPROVED suffix. | Auto-wired on /plugin install |
| Statusline | Status bar showing model, working folder, context-window usage (ctx: 45.2k/200.0k (23% used)), and 5h/7d rate-limit consumption. | Manual setup — see extras/statusline/README.md |
Hooks are installed automatically when you /plugin install this marketplace entry. The statusline is a global settings.json setting and can't be plugin-installed; instructions are in extras/.
/plugin marketplace add KavasiMihaly/AI-plugins
/plugin install claude-code-essentials@OneDayBI-Marketplace
The statusline is a global Claude Code setting, not a plugin-installable hook. See extras/statusline/README.md for copy-paste setup.
/reload-plugins
If hooks don't fire, fully restart Claude Code (hooks snapshot at session start).
_Agent Logs/_Agent Logs/
├── .gitignore # auto-generated, ignores all log files
├── session-2026-05-26-a1b2c3d4.jsonl # one file per session
└── sessions-index.jsonl # summary index
Each JSONL entry looks like:
{"ts":"2026-05-26T14:32:11Z","event":"PreToolUse","session_id":"a1b2c3d4","tool":"Bash","summary":"npm install lodash","details":{"command":"npm install lodash","timeout":null,"background":false}}
The plugin drops a .gitignore inside _Agent Logs/ on first session so logs are ignored by default in every repo.
When Claude tries to run pip install dbt-core, the hook blocks it and prompts:
Install command detected. Please confirm with the user, then append
# APPROVEDto the command.
Claude asks you, you say yes, Claude re-runs as pip install dbt-core # APPROVED — the hook lets it through.
Claude Opus 4.7 | my-project | ctx: 45.2k/200.0k (23% used) | 5h: 12% 7d: 4%
See assets/statusline.png for the rendered version in a real Claude Code session.
.env, credentials, keys, tokens, passwords.<your-repo>/_Agent Logs/. Nothing is sent anywhere.No configuration required for the default experience. Per-tool tweaks (log retention, sensitive-pattern list, etc.) are in hooks/agent-logger.py and hooks/block-installs.py — they're short, readable scripts.
PATH (used by the hooks)jq (only for the statusline shell variant — Python variant is preferred)MIT — see LICENSE.
Mihaly Kavasi — @KavasiMihaly | OneDayBI | Self-Service BI Blog
Found this useful? Star the repo and browse the rest of the OneDayBI Marketplace. For patterns and best practices behind these hooks, see the Claude Code Handbook.
Matches all tools
Hooks run on every tool call, not just specific ones
Executes bash commands
Hook triggers when Bash tool is used
Own this plugin?
Verify ownership to unlock analytics, metadata editing, and a verified badge. GitHub access is read-only (username + org membership).
Sign in to claimOwn this plugin?
Verify ownership to unlock analytics, metadata editing, and a verified badge. GitHub access is read-only (username + org membership).
Sign in to claimBased on adoption, maintenance, documentation, and repository signals. Not a security audit or endorsement.
npx claudepluginhub kavasimihaly/claude-code-essentialsEnd-to-end dbt pipeline automation for SQL Server. CSV to star schema with staging, dimensions, facts, tests, and validation.
End-to-end Power BI Dataflow Gen1 to Microsoft Fabric medallion notebook migration. Export, analyze, and generate bronze/silver notebooks on either engine — distributed PySpark (synapse_pyspark) or single-node Python (polars/duckdb/delta-rs) — then deploy and validate in Fabric.
Harness-native ECC operator layer - 67 agents, 271 skills, 92 legacy command shims, reusable hooks, rules, selective install profiles, and production-ready workflows for Claude Code, Codex, OpenCode, Cursor, and related agent harnesses
Reliable automation, in-depth debugging, and performance analysis in Chrome using Chrome DevTools and Puppeteer
Comprehensive feature development workflow with specialized agents for codebase exploration, architecture design, and quality review
Core skills library for Claude Code: TDD, debugging, collaboration patterns, and proven techniques
Real-time statusline HUD for Claude Code - context health, tool activity, agent tracking, and todo progress
Continuous self-referential AI loops for interactive iterative development, implementing the Ralph Wiggum technique. Run Claude in a while-true loop with the same prompt until task completion.