Getting it into your agent
One page per mod, every tool's command on it. A separate URL per tool would split the same page into five that compete with each other.
npx agentmods add skills/pacphi/agentic-kit/ruflo-token-auditnpx skills add pacphi/agentic-kit --skill ruflo-token-auditgit clone --depth 1 https://github.com/pacphi/agentic-kitWhat it costs to keep this loaded
Counted locally with the o200k_base tokenizer, which is exact for GPT models; Claude uses its own tokenizer and its counts differ. Treat this as one consistent yardstick across the catalogue rather than a bill. Prices are per million input tokens.
| Model | Per session | Once invoked |
|---|---|---|
| Fable 5 | $0.00105 | $0.01366 |
| Opus 5 | $0.00053 | $0.00683 |
| Sonnet 5 | $0.00021 | $0.00273 |
| Haiku 4.5 | $0.00011 | $0.00137 |
Grade A, and why
ruflo-token-audit scanned grade A with 0 findings against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured 2d ago.
A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.
Nothing flagged
None of the 26 patterns this scan looks for appear in this file: no shell pipes, no recursive deletes, no credential paths, no hidden text, no instruction-override or anti-refusal phrasing, no agent-config snooping. That is not a guarantee, it is the absence of the things that are checkable.
How it starts
The opening of the file, as written. The whole thing — 93 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Claude Code Usage Audit (ruflo-token-audit)
A comprehensive picture of Claude Code usage, built from the local session transcripts
in ~/.claude/projects/**/*.jsonl (each assistant message records its token usage, tool
calls, model, and metadata). This is not limited to "tokens ruflo burns" — it covers
ALL Claude Code activity (interactive, subagents, hooks, MCP, web tools) and answers two
questions: where is my usage going? and is any of it runaway automation?
This skill is self-contained — it bundles its own engine, so it works even if the full agentic-kit kit isn't installed.
When to use
Trigger on: "where are my tokens/usage going", "why is my usage so high", "break down my Claude Code activity", "I'm hitting my Max/Pro limit", "what am I spending tokens on", "is the plan worth it". Also proactively if the user mentions surprising usage.
Procedure
-
Run the bundled engine. Prefer the copy that ships inside this skill (works with no kit install); fall back to the PATH command if present:
# self-contained (always available wherever this skill is installed): python3 ~/.claude/skills/ruflo-token-audit/scripts/ruflo-token-audit.py --days 7 # …or, if the agentic-kit kit put it on PATH: ruflo-token-audit --days 7- Honor any window the user gives ("past month" →
--days 30). --top Nwidens each section;--jsongives machine-readable output;--no-daemonsskips thepscross-reference.- If
python3isn't found, say so — the engine is stdlib-only Python 3.
- Honor any window the user gives ("past month" →
-
Read the whole picture, then lead with the headline. The report has many sections; synthesize, don't echo. Key sections and what they tell you:
Section Read it for BY MODEL Opus = interactive; heavy Haiku/Sonnet = automation/subagents SESSIONS PER DAY tens = human; hundreds–thousands = automation (≈one/min = robotic) ACTIVITY BY HOUR a flat 24h histogram (busy at 3am) is automation, not a person TOOL USAGE what the work actually is (Bash/Read/Edit vs Task/MCP) MCP USAGE per-server call volume; heavy MCP also means big per-session tool-def tax SUBAGENT FAN-OUT Task spawns + sidechain share — how much is delegated/parallel BUSIEST SESSIONS a single runaway conversation surfaces here by token total CACHE EFFICIENCY high cache-read% is normal/cheap; flag only with huge automated volume STARTUP CONTEXT TAX fixed per-session cost (CLAUDE.md + tool/skill manifests) × many sessions RUNNING DAEMONS live ruflo daemon startmapped to top-burn projects (the classic leak)
What ships with it
1 file beside SKILL.md in the same directory: the scripts, references and assets a skill reads on demand. Not counted in the per-session cost; read them before you install if any of them is executable.
What this file has done since we first saw it
Hashed on every crawl. A supply-chain change to an agent config is a question of when, not whether, so the history is kept rather than the latest state alone.
- 2d ago First seen · 93 lines · 105 tokens per session scan A d126d82e8717
ruflo-token-audit is a skill published in the GitHub repository pacphi/agentic-kit (49 stars, last pushed 3d ago), licensed MIT. It adds 105 tokens to every session and 1,366 once invoked, about $0.0005 per session on Opus 5. A static security scan graded it A with 0 findings. No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-30.
Other skills, from other repositories
debugging-output-and-previewing-html-using-ray
Use when user says "send to Ray," "show in Ray," "debug in Ray," "log to Ray," "display in Ray," or wants to visualize data, debug output, or show diagrams in the Ray desktop application.
fortify-development
ACTIVATE when the user works on authentication in Laravel. This includes login, registration, password reset, email verification, two-factor authentication (2FA/TOTP/QR codes/recovery codes), passkeys, profile updates, password confirmation, or any auth-related routes and controllers. Activate when the user mentions…
Shade dropdown surface contract
DropdownMenu, Select, and Popover share one visual recipe (bg-surface-elevated-2 + border-border/60 dark:border-border/30 + shadow-md). Change them together. Trigger when editing any of those three Shade files.
Shade ShadCN install
Guardrails for running pnpm dlx shadcn@latest add in Shade — never overwrite existing components, fresh branch first, swap raw colours for semantic tokens after integrating. Trigger when the user proposes a shadcn add, or when a fresh ShadCN-shaped file lands in apps/shade/src/components/ui.
Shade component decision
Decide which Shade layer (Token, Primitive, Component, Recipe, Pattern) a new piece of UI belongs in — and whether it should be added to Shade at all. Trigger when creating new files in apps/shade/src/components, or proposing to add a new Shade component.
tinybird-cli-guidelines
Tinybird CLI commands, workflows, and operations. Use when running tb commands, managing local development, deploying, or working with data operations.