Wrap long-running commands with per-service runners and a Unix socket HTTP API.
npx skills add https://github.com/bbatchelder/tap --skill tap
Please help me install this Agent Skill. Skill: tap Source: https://github.com/bbatchelder/tap/tree/main Command: npx skills add https://github.com/bbatchelder/tap --skill tap
Tap solves the challenge of reliably starting, monitoring, and controlling long-running processes by wrapping them with a dedicated runner that captures output in memory and exposes a minimal API to query logs and manage restarts.
Use tap to run a service, observe logs, and restart when needed. Example: tap run myapp -- node server.js; then tap observe myapp --last 50; if needed, tap restart myapp.
High-intent search queries and answers about installing and using this skill.
To supervise a long-running command with queryable logs, wrap it with a per-service runner that captures stdout and stderr in an in-memory ring buffer. This isolates process lifecycle and enables fast log access without writing log files.
Remote control of a long-running process is handled via a lightweight HTTP API over Unix-domain sockets. You can send restart or stop commands to a specific service runner without direct terminal access.
Yes, per-service isolation assigns a dedicated runner to each long-running command on Unix-like environments. This enables predictable restarts, individual readiness checks, and independent log buffers for each supervised process.
You can observe recent output from a supervised process by querying the in-memory ring buffer using the CLI. Running the observe command retrieves the latest stdout and stderr lines directly from memory.
An in-memory ring buffer for process supervision limits log history to recent lines and loses all output upon process exit or system restart. It does not provide persistent log files for long-term auditing or crash analysis.
myaiartist360-makerManage persistent tmux sessions and capture pane output for terminal workflows.
wojonsIntegrate logs with metrics, traces, alerts, and dashboards for system observability.
forwardimpactSupervise daemon processes and manage service lifecycles via Unix sockets.
cofinImplement zero-copy IPC with shared memory and ring buffers in Rust.
