Jane-o-O-o-O / agent-pulse-skill

The SKILL.md maps what a user asks for onto the underlying CLI's own subcommands: "recent usage" to status and top, "models and cost" to models, leaderboard, optimize, and budget, "risk and health" to alerts, health, anomaly, forecast, and score, and so on through history, sharing, and setup diagnosis. A bundled helper script, run_agent_pulse_snapshot.py, runs a chunk of those subcommands together into one JSON-friendly snapshot, with each subcommand given its own timeout so one slow local scan doesn't block the rest.

The thirteen source platforms it reads from are Hermes, Claude Code, Codex, DeepSeek, OpenClaw, Copilot, Aider, Qwen, OpenCode, Goose, Cursor, Antigravity, and Amp. This is for someone who already runs one or more of those locally and wants an agent to summarize its own session activity, model cost, or health without leaving the chat — the GitHub repository alone does nothing; pip install agentpulse-cli has to happen first, and the skill only wraps a tool it does not ship.

READ THE FULL ANALYSIS＋

No license is declared. Neither the repository nor the individual skill record in our data carries a license field, which is worth checking directly before relying on this beyond personal, local use.

One more listing exists for this repository, but it is dead. A second row, a bare placeholder titled "Jane O O O O Skills" with the description "More skills from Jane O O O O," carries no skill content of its own and is not counted or described above.