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/auditgit clone --depth 1 https://github.com/tacoda/keystoneWhat 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 | $0.00007 | $0.00786 |
| Opus 5 | $0.00003 | $0.00393 |
| Sonnet 5 | $0.00001 | $0.00157 |
| Haiku 4.5 | $0.00001 | $0.00079 |
Grade A, and why
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 — 47 lines — stays where its author put it; the contents beside it link to each section on GitHub.
audit
Full dual-flywheel audit. Learning (capture from review) and Pruning (remove dead rules). Periodic. Read .charter/README.md (flywheels section) and .charter/learning/README.md.
Learning flywheel
- Walk
.charter/learning/inbox/and flag candidates that should be promoted. - Run review sensors on recent commits — surface patterns that should become guides.
- Capture findings as new inbox entries (see
learn.md).
Pruning flywheel
The Pruning flywheel reads the charter-debt sensor and proposes the diff to .charter/corpus/state/charter-debt.md. Categories the sensor surfaces:
- Stale rules — guides in
.charter/guides/that no diff has touched in N months. - Dead idioms — idiom dirs whose stack is no longer in
CODEBASE_STATE.md. - Placeholders — bootstrap
<...>markers left unfilled. - Failing sensors — sensors recorded as available that error on invocation.
- Empty shells — scaffolded dirs with no real content.
- Uncited policies — installed policies whose guides were never referenced.
- Unresolved gaps —
.charter/adapters/<agent>/<topic>.mdTODO placeholders. - Drifted state — stack-drift findings +
CODEBASE_STATE.mdstale-last_reconciled. - Strict-cascade violations — run
keystone policy verify; any strict-violation output is a hard debt entry. A project file at.charter/<kind>/<name>.mdis overriding astrictitem from a team or org policy. Remove the shadowing file or escalate to the policy owner. - Required-item gaps — same
keystone policy verifyrun. Required items with no definition anywhere in the cascade are advisory debt: the project is expected to define them at.charter/<kind>/<name>.md. Pruning surfaces them; the project decides which to fill in.
Then update the empirical state files:
- Risk fingerprint — read
.charter/sensors/risk-fingerprint.mdand update.charter/corpus/state/risk-fingerprints.md. - Traffic topology — read
.charter/sensors/traffic-topology.mdand update.charter/corpus/state/traffic-topology.md.
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 · 47 lines · 7 tokens per session scan A ee2132e6a64f
audit is a command published in the GitHub repository tacoda/keystone (44 stars, last pushed 2mo ago), licensed MIT. It adds 7 tokens to every session and 786 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
git
Git operations with intelligent commit messages and workflow optimization.
checklist
Generate a custom checklist for the current feature based on user requirements.
clarify
Identify underspecified areas in the current feature spec by asking up to 5 highly targeted clarification questions and encoding answers back into the spec.
specify
Create or update the feature specification from a natural language feature description.
analyze
Perform a non-destructive cross-artifact consistency and quality analysis across spec.md, plan.md, and tasks.md after task generation.
converge
Assess the current codebase against the feature's spec, plan, and tasks, then append any remaining unbuilt work as new tasks to tasks.md so implement can complete it.