usage-analysislisted

Use when the user asks an interpretive question about their Claude Code token spend — what's interesting, what's anomalous, what to change. This skill produces insights and questions; it does NOT enumerate top consumers (the dashboard does that). Use `usage-dashboard` for "show me the numbers." Use `claude-audit` for "audit my config overlap." Use this skill for "tell me what's interesting / what should I change." Trigger phrases: "am I close to my Claude limit", "how much Sonnet am I using", "where are my Claude tokens going", "what's eating my Claude budget", "why is my Claude spend so high", "what should I change about my Claude setup".
glitchwerks/claude-prospector · ★ 1 · Code & Development · score 67

Install: claude install-skill glitchwerks/claude-prospector

# Usage Analysis Skill You are surfacing **insights and improvement suggestions** about the user's Claude Code token spend that are not obvious from the dashboard at a glance. Your output is observations + questions, not a stats rehash. ## Companion skills — route the user away when appropriate | User wants | Skill | | --- | --- | | Show me my numbers / regenerate the dashboard | `usage-dashboard` | | Audit my agent/skill config for overlap or conflicts | `claude-audit` | | Tell me what's interesting / what should I change | **this skill** | ## Prerequisites This skill invokes `python -m claude_prospector` under the hood. The Python package must be installed in the environment Claude Code uses. See the [README install steps](https://github.com/glitchwerks/claude-prospector#install-as-a-claude-code-plugin) for the two-step install. ## Input — structured JSON, not the rendered dashboard Get the structured payload directly from the CLI. Do **not** screen-scrape the dashboard HTML. ```bash python -m claude_prospector dashboard --format json --no-open ``` `--format json` writes the same `DATA` object the dashboard embeds, to stdout. Progress lines go to stderr. Pipe through `jq` (or read into memory) and reason over it. Add `--window 7d` / `--window 5h` / `--from --to` to scope. Run multiple windows if comparing — e.g. 7d vs. all-time to see whether a category is growing or shrinking. Skill-tracking events (passed vs. invoked) live in per-day JSONL files at: - POSIX: `