session-reportlisted

Use when the user asks what they are spending on Claude, how many tokens something cost, why a session got expensive, or how well caching is working. Triggers: 'session report', 'token usage', 'how many tokens', 'what did that cost', 'am I burning tokens', 'cache hit rate', 'why is this session so expensive', 'which project costs most', 'show my usage', 'usage this week', 'last 7 days', 'claude stats', 'token breakdown', 'subagent cost'. Also use proactively after a long or context-heavy session when the user wonders where the budget went. Reads Claude Code's own JSONL transcripts under ~/.claude/projects; on other harnesses it reports that no compatible transcripts exist rather than inventing numbers.
Tamircohen28/tamirs-superpowers · ★ 2 · Data & Documents · score 74

Install: claude install-skill Tamircohen28/tamirs-superpowers

# Session Report ## Compatibility — this skill is Claude Code specific **It is not portable, and it does not pretend to be.** `analyze-sessions.mjs` parses Claude Code's own JSONL transcript format, and it depends on details of that format that no other harness produces: - one API response split across multiple `type:"assistant"` entries sharing a `requestId`, where only the last carries the final `output_tokens` (dedupe by `requestId`, keep the max — without this the totals are 3–10× too high); - `isSidechain` / `isMeta` / `isCompactSummary` flags to tell a human turn from a tool result, an interrupt marker, or a compaction summary; - subagent transcripts at `<project>/<sessionId>/subagents/*.jsonl` with a sibling `*.meta.json` carrying `agentType`; - `usage.cache_read_input_tokens` / `cache_creation_input_tokens` for the cache-hit rate; - global dedupe by entry `uuid`, because a resumed session re-serializes prior entries. These were established empirically against Claude Code transcripts. They are Claude Code internals, not a standard. | Target | Status | Basis | |---|---|---| | Claude Code | supported | Writes the format this parser reads | | Claude Desktop | partial | Same plugin artifact, but its transcripts are only readable here if they land in `~/.claude/projects` in the same JSONL shape. The skill **checks this at run time** (step 0a) rather than assuming it. | | Codex CLI, Cursor, Gemini CLI, OpenCode | unsupported | None of them writes this format. Th