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 commands/tacoda/keystone/check-driftgit clone --depth 1 https://github.com/tacoda/keystoneWrote 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/tacoda/keystone/check-drift)<a href="https://agentmods.dev/commands/tacoda/keystone/check-drift"><img src="https://agentmods.dev/badge/commands/tacoda/keystone/check-drift.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 | $0.00009 | $0.00397 |
| Opus 5 | $0.00005 | $0.00198 |
| Sonnet 5 | $0.00002 | $0.00079 |
| Haiku 4.5 | $0.00001 | $0.00040 |
Grade A, and why
check-drift 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 3d 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
check-drift
Run the drift sensor on the current diff. Compare the in-progress changes against loaded charter rules. Fast pre-verify check. Read .charter/sensors/drift.md and .charter/guides/process/implementation.md.
Activities
- Identify the diff —
git diff, the staged changes, or the in-progress edits. The list of changed files is the touched-files set for this action. - List which guides apply — load
.charter/guides/idioms/<stack>/*.mdfor each touched stack, plus.charter/guides/process/implementation.mdand any guides from installed policys (.charter/policies/*/guides/). For each candidate, if it declaresglobs:in frontmatter, keep it only when at least one touched file matches; guides withoutglobs:keep today's loading. - Compare diff to guides — line by line, surface violations or hot drift signals. A guide's findings are reported only against files matching its
globs:(or all touched files when noglobs:is set). - Report findings — one bullet per drift. Cite the guide that was violated.
When to invoke
Between implementation and verify. Faster and cheaper than the full verify pass; meant to catch obvious drifts before sensors run.
Index freshness precondition
check-drift reads .keystone/INDEX.json to know which guides apply to the touched files. If the index is stale (any guide file modified more recently than INDEX.json), run keystone index first — otherwise the drift check uses an outdated set of rules.
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.
- 3d ago First seen · 24 lines · 9 tokens per session scan A 9600cb61421d
check-drift is a command published in the GitHub repository tacoda/keystone (44 stars, last pushed 2mo ago), licensed MIT. It adds 9 tokens to every session and 397 once invoked, about $0.0000 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
observe
部屋を能動的に観察する。3つの基本ブロックを自由に選びながらループし、受動的なスナップショットではなく発見のある観察を行う。.
model
Check if an AI model fits on your GPU before downloading.
cuda-install
Get step-by-step CUDA Toolkit installation instructions tailored to your platform.
eol-report
EOL/EOS risk report — devices, OS versions, and firmware approaching or past end-of-life/end-of-support, prioritized by criticality.
gum-search
BM25 search the embedded catalog (TTY table, pipe JSON).
rvcsi-calibrate
Learn a v0 per-subcarrier baseline (mean amplitude) from a .rvcsi capture.