メニュー
Ultra-fast token+cost computation for Codex + Claude SSE logs (Rust) plus Python anchor/fast comparisons; includes build+run+benchmark guide.
Codex または Claude でインストール この Prompt をコピーして Codex、Claude、または他のアシスタントに貼り付けると、Skill ページを確認してインストールできます。
SOC 職業分類に基づく
| name | compute-token-stats-rust-ultra |
| description | Ultra-fast token+cost computation for Codex + Claude SSE logs (Rust) plus Python anchor/fast comparisons; includes build+run+benchmark guide. |
[Created by Codex: 019beb87-b904-7a81-a7c9-5b5d5575d2f7 2026-01-23] [Edited by Claude: 9d1228e9-aabe-4ab3-9b9c-c0c1d1f64ebc 2026-01-27]
If the user asked you to load this skill, you MUST use it for token cost computation.
# WRONG - will hang waiting for stdin forever!
/tmp/.claude/skills/compute-token-stats-rust-ultra/rust/target/release/token_costs
# CORRECT - always pipe data to it
cat file.jsonl | /tmp/.claude/skills/compute-token-stats-rust-ultra/rust/target/release/token_costs
To compute tokens for specific sessions, filter first then pipe:
# Filter by single SID
grep '"sid":"abc123"' ~/centralized-logs/claude/sse_lines.jsonl | \
/tmp/.claude/skills/compute-token-stats-rust-ultra/rust/target/release/claude_token_costs
# Filter by multiple SIDs (use grep -F with file)
grep -F -f /tmp/sids.txt ~/centralized-logs/codex/sse_lines.jsonl | \
/tmp/.claude/skills/compute-token-stats-rust-ultra/rust/target/release/token_costs
Compute token totals + estimated costs from massive JSONL SSE logs:
~/centralized-logs/codex/sse_lines.jsonl~/centralized-logs/claude/sse_lines.jsonlThis skill includes:
rust/ → token_costs + claude_token_costspython/run_compute_token_costs_anchor.py → runs ~/swe/token-computation/compute_token_costs.pypython/compute_{codex,claude}_token_costs_fast_pipe.pycd /tmp/.claude/skills/compute-token-stats-rust-ultra/rust && \
cargo build --release --bins
Codex:
cat ~/centralized-logs/codex/sse_lines.jsonl | \
/tmp/.claude/skills/compute-token-stats-rust-ultra/rust/target/release/token_costs
Claude:
cat ~/centralized-logs/claude/sse_lines.jsonl | \
/tmp/.claude/skills/compute-token-stats-rust-ultra/rust/target/release/claude_token_costs
Generate a pricing snapshot (recommended; makes runs reproducible):
python /Users/sotola/swe/token-computation/pull_pricing.py \
--providers openai,anthropic --timeout 20 --snapshot-dir ~/centralized-logs/pull-pricing/snapshots
Then run:
python /tmp/.claude/skills/compute-token-stats-rust-ultra/python/run_compute_token_costs_anchor.py \
--codex ~/centralized-logs/codex/sse_lines.jsonl \
--claude ~/centralized-logs/claude/sse_lines.jsonl \
--pricing-snapshot ~/centralized-logs/pull-pricing/snapshots/<pick-one>.json
This compares Rust Ultra (streaming stdin) vs Python anchor on the same datasets and prints:
python /tmp/.claude/skills/compute-token-stats-rust-ultra/python/bench_rust_vs_anchor.py
When turn.session_configured is missing for some sid (log truncation / invalid state machine), per-sid model attribution is unknowable.
compute_token_costs.py) prices Codex totals with a single detected codex.model.(unknown) it prices using the detected pricing model to stay consistent with the anchor.Shadow and version-control another agent's conversation. Use when asked to track an agent, create agent biography, shadow an agent, document agent work, or make agent output more accessible.
Configure Agent HQ Electron crash mitigations + diagnostics via repo .env (added after 2026-01-22 renderer crash).
Start the full agent stack (tmux sessions for Claude Code agents, SQLite ingestors, announcers, watchdog, and mitmproxy) after a fresh Mac reboot. This skill should be used when the user asks to "start everything", "bring up the agent stack", "start after reboot", "launch tmux sessions", or needs to initialize the development environment with all monitoring and agent processes.
Bootstrap the complete agentic ecosystem (Claude Code, Codex, vscode-shims, agent-box) on a remote machine. Use when user asks to install ecosystem, setup remote agent, bootstrap agent box, deploy to remote, or install claude/codex on another machine.
This skill should be used when containerizing a multi-service agentic ecosystem (UI + Agent Management Server, optionally launchers/ingestors) with Docker Compose while preserving localhost-only security, correct in-container networking, centralized log volume mounts, and working Codex/Claude CLIs.
Safe incremental deployment and updates for the agentic ecosystem after initial setup. Focuses on config preservation, safe service restarts, verification, and optional testing. Use after agentic-ecosystem-remote-deployment has completed at least once.