Borrowing it
Nothing to install: this file belongs to adamw7/tools. Take a copy, put it at the same path in your own repository, and replace the rules that are about this project with yours.
curl -O https://raw.githubusercontent.com/adamw7/tools/main/.claude/agents/doc-drift-auditor.mdgit clone --depth 1 https://github.com/adamw7/toolsWrote 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/adamw7/tools/doc-drift-auditor)<a href="https://agentmods.dev/agents/adamw7/tools/doc-drift-auditor"><img src="https://agentmods.dev/badge/agents/adamw7/tools/doc-drift-auditor/github.svg" alt="Measured on agentmods" height="20"></a>Or the 80×15 button, for a site that already has a row of RSS and ATOM ones. Only the verdict fits; the numbers stay here.
<a href="https://agentmods.dev/agents/adamw7/tools/doc-drift-auditor"><img src="https://agentmods.dev/badge/agents/adamw7/tools/doc-drift-auditor.svg" alt="Reviewed on agentmods" width="80" 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 | $0.00072 | $0.01010 |
| Opus 5.5 | $0.00029 | $0.00404 |
| Sonnet 5.5 | $0.00014 | $0.00202 |
| Haiku 4.5 | $0.00007 | $0.00101 |
Grade A, and why
doc-drift-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 1mo 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 — 78 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Doc Drift Auditor
The documentation here is a build artifact, but the build only pins what a rule was configured to pin. Everything else — how many skills there are, what a catalogue table lists, which commands a section recommends — drifts silently and is found by a reader, usually an agent acting on a stale instruction.
Read the documents against each other and against the code, and report what no longer agrees. Report only: leave the fixing to the caller, who knows which side is the truth.
What the build already pins — do not re-report it
These have a rule, so a mismatch fails the build and is not drift:
crossDocConsistency— the Java version, captured byJava (\d+)in bothCLAUDE.mdandAGENTS.md.readmeConsistency— the protobuf major version,proto(\d), betweenREADME.mdandAGENTS.md.moduleMapConsistency— every<module>of the root pom is mentioned inCLAUDE.mdandAGENTS.md. Presence only: that a module is named proves nothing about what the sentence next to it claims.skillFilesExist,uniqueNames,uniqueDescriptions— each skill, sub-agent and command has a well-formed definition with a name and description nobody else uses.contextBudget—CLAUDE.mdfits 32 KB.
Where drift actually lives
Work through these, cheapest first:
- Counts written as words.
grepforthirteen,three,four,five,nineand their neighbours inCLAUDE.mdandAGENTS.md, then count the thing on disk: skill directories under.claude/skills, sub-agents under.claude/agents, commands under.claude/commands, MCP servers (aMain.javaunder anmcppackage), modules in the root pom. A count is the first thing to rot and the cheapest to check. - Catalogue tables against the code.
AGENTS.md's rule catalogue must have a row per@Namedrule inclaude-code-enforcer; its skill table a row per skill directory. A rule or skill that exists and is not in the table is invisible to a reader, and a row for something deleted is worse. - The rules the profile wires. Compare the
<rules>block of the root pom'sclaude-md-enforceprofile against the catalogue's claims about what is wired, optional, or deliberately left out. - Commands that are quoted. Every
mvninvocation in the documents should name a profile, module or property the poms actually define — a-Pthat no pom declares runs and quietly does nothing. - File references. Links and paths named in prose must exist.
claudeMdFormatcan check this whenvalidateFileReferencesis on; verify how it is configured before assuming a broken path would have failed the build. - Sections that describe the same thing twice.
CLAUDE.mdsummarises whatAGENTS.mdstates in full. Where the summary says something the full text no longer does,AGENTS.mdwins — that is the stated contract — so report it as the summary being stale.
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.
- 1mo ago First seen · 78 lines · 72 tokens per session scan A 965a7d7bf1ea
doc-drift-auditor is an agent published in the GitHub repository adamw7/tools (11 stars, last pushed today), licensed MIT. It adds 72 tokens to every session and 1,010 once invoked, about $0.0003 per session on Opus 5.5. 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 agents, from other repositories
bash
A command-line agent for running Bash shell commands, terminal tasks, and Git operations. Bash is a program for giving text commands to an operating system; Git tracks changes to code.
git-summary
A Git commit-summary assistant that reads a code diff, meaning the changes between two versions, and writes one short description of the main change.
git-commit-assistant
A Git commit agent that splits workspace changes into focused commits and writes Chinese Conventional Commit messages in a required dependency order.
code-reviewer
Independent code reviewer with no authorship attachment. Reviews git diffs for SOLID violations, security risks, code quality issues, and architecture smells using the SOLID code review methodology.
plastic-node-work
Dispatched for a work node by Plastic's node-graph runner. Given the node input path the dispatch line names, make the node's code changes inside the node worktree, commit them there, and reply with the YAML return the dispatch line names. Never spawned outside a runner dispatch.
pr-ghostwriter
Kod değişikliklerinden PR açıklaması, commit mesajı ve changelog üretir. Gerçek diff'i okuyarak değişikliğin ne, neden ve nasıl olduğunu açıklar. Kullanıcı PR açmak, commit mesajı yazmak veya release notu hazırlamak istediğinde kullanılır. Jenerik açıklama üretmez — her zaman gerçek değişikliğe özgü yazar.