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.
git clone --depth 1 https://github.com/vukkt/token-wardenWrote this? Show the measurements
A badge with what this costs and how it scanned, read live from this page, so it follows the numbers instead of freezing them. Markdown for a README, HTML for a documentation site or a project page.
[](https://agentmods.dev/commands/vukkt/token-warden/warden-power)<a href="https://agentmods.dev/commands/vukkt/token-warden/warden-power"><img src="https://agentmods.dev/badge/commands/vukkt/token-warden/warden-power.svg" alt="Measured on agentmods" height="20"></a>What 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 5.1 | $0.00053 | $0.00394 |
| Opus 5 | $0.00026 | $0.00197 |
| Sonnet 5 | $0.00011 | $0.00079 |
| Haiku 4.5 | $0.00005 | $0.00039 |
Grade A, and why
warden-power scanned grade A with 0 findings against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured 8d ago.
A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.
Nothing flagged
None of the 26 patterns this scan looks for appear in this file: no shell pipes, no recursive deletes, no credential paths, no hidden text, no instruction-override or anti-refusal phrasing, no agent-config snooping. That is not a guarantee, it is the absence of the things that are checkable.
What it actually says
Run the token-warden power command:
cd "${CLAUDE_SKILL_DIR}/.." && npx tsx src/power.ts $ARGUMENTS
Per agent it reports, from the recorded golden-replicate variance:
- A table of runs-per-side (2, 3, 5, 8, 12) against the standard error and the minimum detectable saving at 80% and 90% power — the smallest true saving the 2x-rent + z-SE gate would actually promote at that budget.
- With
--target-saving N: the required runs per side to detect that saving at 80% and 90% power. A target at or below the 2x-rent bar is flagged as undetectable at any run count — the gate itself rejects it. - With
--runs N: the MDS at exactly that n, plus (with a target) the achieved power percentage. --rent Noverrides the default rent (median context cost of the agent's active rules, 25 when none are deployed).
The estimate is conservative: it assumes uniform run allocation, and the selector's Neyman variance-proportional top-up only tightens the SE. Zero tokens, read-only. Relay the command output to the user.
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.
- 8d ago First seen · 30 lines · 53 tokens per session scan A 414e0e3c01b7
warden-power is a command published in the GitHub repository vukkt/token-warden (13 stars, last pushed today), licensed MIT. It adds 53 tokens to every session and 394 once invoked, about $0.0003 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 commands, from other repositories
test-cards
Display Stripe test card numbers for various testing scenarios.
memory-report
Calibration + inventory check of your causal memory (prediction ledger accuracy, what you know).
recall
Recall past experience relevant to a task or decision before acting on it.
test
Run Postman collection tests, analyze results, diagnose failures, and suggest fixes.
tree-ring-certify
Generate Tree Ring harness or recall-quality evidence without confusing it with the full framework release suite.
debug
Systematic root-cause investigation, then fix with TDD — never proposes a fix before finding the cause; resumable diagnosis + regression test in .orc/ state. Workspace-aware. For a quick one-off look without session state, prefer the bundled /debug.