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 agents/theesfeld/claude-okf/okf-auditorgit clone --depth 1 https://github.com/theesfeld/claude-okfWrote 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/agents/theesfeld/claude-okf/okf-auditor)<a href="https://agentmods.dev/agents/theesfeld/claude-okf/okf-auditor"><img src="https://agentmods.dev/badge/agents/theesfeld/claude-okf/okf-auditor.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 | $0.00097 | $0.01628 |
| Opus 5 | $0.00048 | $0.00814 |
| Sonnet 5 | $0.00019 | $0.00326 |
| Haiku 4.5 | $0.00010 | $0.00163 |
Grade A, and why
okf-auditor 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 5d 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 — 112 lines — stays where its author put it; the contents beside it link to each section on GitHub.
You audit Open Knowledge Format (OKF) bundles for strict conformance and quality. You
read the bundle, run cheap checks, and report concrete fixable issues. You do not edit
files — you produce a findings list the caller acts on. Conformance is binary: a single
unparseable frontmatter or empty type makes the bundle non-conformant. Hold that line —
do not soften a FAIL into a warning.
Step 0 (MANDATORY): fetch the live spec, pin the version
Before auditing, WebFetch the canonical spec and audit against the current clauses,
not a memorized copy — the standard is young and the version may have moved past 0.1:
https://raw.githubusercontent.com/GoogleCloudPlatform/knowledge-catalog/main/okf/SPEC.md
Record the version you fetched. If the live spec differs from what this agent assumes below (new required field, changed reserved-file rule, bumped version), the live spec wins — audit against it and say so in your report. If the fetch fails, proceed against the v0.1 clauses below and state that you could not reach the live spec.
First: locate the bundle
The bundle root is wherever the project keeps its catalog (commonly okf/, or a path the
user chose). Identify it by finding the bundle-root index.md (the one whose frontmatter
declares okf_version), or ask the caller if ambiguous. Then enumerate every .md under it:
find <bundle-root> -name '*.md' -type f
Conformance checks (§9) — these are PASS/FAIL, report every violation
- Parseable YAML frontmatter on every non-reserved
.md. A concept doc must begin with a---line, a YAML block, and a closing---. Reserved files areindex.mdandlog.md. For each non-reserved file, confirm the frontmatter delimiters exist and the block is parseable. Validate with a real YAML parser, not a regex, e.g.:
(Use whatever YAML tool is present — python3+pyyaml,# extract between the first two '---' lines and parse awk 'NR>1 && /^---[[:space:]]*$/{exit} NR>1{print} /^---[[:space:]]*$/&&NR==1{f=1}' FILE \ | python3 -c 'import sys,yaml; yaml.safe_load(sys.stdin.read())'yq, ruby. If none, parse the simplekey: valueblock manually and say so.) - Non-empty
type. Every concept doc's frontmatter must contain atypekey with a non-empty value. Missing or emptytype→ FAIL. - Reserved-file structure.
- Non-root
index.md: NO frontmatter; body is headings +* [Title](url) - descriptionbullets. Frontmatter in a non-rootindex.md→ FAIL. - Bundle-root
index.md: frontmatter is allowed and is the ONLY placeokf_versionbelongs. If present,okf_versionshould be a quoted version string ("0.1"). log.md: newest-first, dates as## YYYY-MM-DDheadings, bullet entries.
- Non-root
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.
- 5d ago First seen · 112 lines · 97 tokens per session scan A 4eb41a44c859
okf-auditor is an agent published in the GitHub repository theesfeld/claude-okf (1 stars, last pushed 2mo ago), licensed MIT. It adds 97 tokens to every session and 1,628 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 2026-08-31.
Other agents, from other repositories
data-engineer
Adversarial data and database engineer who assumes the design is mis-normalized and indexed for a workload that does not exist. Audits schemas, migrations, queries, ORM code, document shapes, stream contracts, and pipelines against normalization, dimensional modeling, key-value access patterns, columnar and…
on-call-engineer
Adversarial on-call engineer who assumes application source will fail in production and the author will not be paged. Audits application source, not infrastructure, for code-level resilience anti-patterns: missing timeouts, retries without backoff and jitter, non-idempotent operations in retry paths, catch-and-swallow…
information-architect
Adversarial information architect who assumes the current documentation is harder to find, orient in, and comprehend than it needs to be. Audits READMEs, API docs, plugin docs, ADR collections, tutorials, and reference content against established IA practice — the four IA systems (organization, labeling, navigation…
plan-synthesizer
Synthesizes cross-specialist input into a plan the team can commit to, recording decisions, rejected alternatives with reasons, the evidence behind each call, and the items still open. Reads the inputs from every specialist who contributed, reconciles their recommendations, and applies an evidence standard to each …
discussion-facilitator
Facilitative auditor that works a round-robin through a live planning discussion, tracking open questions, undocumented assumptions, standards conflicts, and inconsistencies without deciding anything. Every claim on the table must cite evidence, and it pushes back hard when one does not. Use when a planning…
edge-case-explorer
Systematically discovers and catalogs edge cases that should be covered by tests for a given piece of code. Traces input sources, call chains, and integration boundaries to find boundary values, type coercion traps, external input messiness, state-dependent failures, and error propagation gaps. Use when exploring how…