Getting it into your agent
This one installs as part of its plugin. Adding the marketplace and installing the plugin brings it with everything else the plugin ships.
/plugin marketplace add wan-huiyan/claude-ecosystem-hygiene/plugin install doc-freshness-reverse-lintWrote 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/skills/wan-huiyan/claude-ecosystem-hygiene/doc-freshness-reverse-lint)<a href="https://agentmods.dev/skills/wan-huiyan/claude-ecosystem-hygiene/doc-freshness-reverse-lint"><img src="https://agentmods.dev/badge/skills/wan-huiyan/claude-ecosystem-hygiene/doc-freshness-reverse-lint/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/skills/wan-huiyan/claude-ecosystem-hygiene/doc-freshness-reverse-lint"><img src="https://agentmods.dev/badge/skills/wan-huiyan/claude-ecosystem-hygiene/doc-freshness-reverse-lint.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 | $0.00376 | $0.04674 |
| Opus 5 | $0.00188 | $0.02337 |
| Sonnet 5 | $0.00075 | $0.00935 |
| Haiku 4.5 | $0.00038 | $0.00467 |
Grade B, and why
doc-freshness-reverse-lint scanned grade B with 1 finding against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured 11d ago.
A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.
Reads agent configuration directoriesmediumAgent snooping
.claude/, .codex/, .gemini/ hold keys, settings and other credentials a mod has no legitimate need for.
[ -f "$s" ] || s="$(find -L "$HOME/.claude/plugins/cache" -mindepth 5 -maxdepth 5 \ How it starts
The opening of the file, as written. The whole thing — 360 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Doc Freshness Reverse-Lint
Problem
Huiyan's project docs under docs/{research,decisions,findings,runbooks}/ contain normative
guidance. When she later corrects that guidance in ~/.claude/lessons.md or a
feedback_*.md entry, the original project docs stay stale. Future Claude sessions
read them as authoritative and repeat the retracted advice.
Real example: the-causal-impact-repo/docs/research/pre_period_length_methodology.md:92,98
referenced "sorting by p-value" after the user had already decided that approach was wrong.
And a second shape the grep above cannot see (added v1.3.0). A rule is not always contradicted — sometimes it is retired, and the docs that carried it keep a scoped survivor: "X is retired — but it still applies if you are a Y." That reads as current guidance, and it is dead the moment no Y exists.
Real example (DoodleRun, 2026-08-04): "no deploys by agents" was retired across four live briefs, and every retirement kept "a cloud session still cannot" as a live branch — after cloud sessions had stopped existing. The next reader would have paused to work out which kind of session it was instead of just deploying. The retirement was correct; the surviving condition was the defect. The owner caught it by hand, one turn after the retirement shipped.
Mechanisms
1. Reverse-lint (primary, event-driven)
Runs when a negation rule is added to a memory file. Extracts the "don't X" phrase, greps project docs, lists matches. Does NOT edit.
1b. Dead-branch detection (v1.3.0, runs with the reverse-lint)
Same trigger, different question. Where mechanism 1 asks "does any doc still assert the thing you just forbade?", this asks "does any doc still carry the CONDITION of a rule you just retired?"
THREE explicit signals, ANDed — a scoped survivor is often perfectly legitimate, and only a human knows whether the branch can still fire:
- a retirement trigger in the memory file (
retired,superseded,no longer applies,is dead, …); - a quoted subject within 160 chars either side of it — retirement prose
nearly always quotes the rule it retires (
"...",'...',`...`,**...**), and the quote must be ≥ 2 tokens and not a path or filename; - a doc paragraph containing that subject with a surviving-conditional
marker after it (
still cannot,only applies,unless you,except when,if you are a,remains true for, …).
What ships with it
10 files beside SKILL.md in the same directory: the scripts, references and assets a skill reads on demand. Not counted in the per-session cost; read them before you install if any of them is executable.
- .claude-plugin/plugin.json 812 B
- evals/fixtures/fake_lesson_pvalue.md 310 B
- evals/fixtures/project_with_dead_branch/docs/runbooks/deploy_thing.md 184 B
- evals/fixtures/project_with_dead_branch/docs/runbooks/unrelated.md 228 B
- evals/fixtures/project_with_stale_doc/docs/research/pre_period_stale.md 169 B
- evals/validation_case.md 2.5 KB
- README.md 5.3 KB
- scripts/hook_dispatch.sh 3.3 KB runs code
- scripts/reverse_lint.py 18 KB runs code
- scripts/weekly_audit.py 5.7 KB runs code
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.
- 11d ago First seen · 360 lines · 376 tokens per session scan B db264a3bfa3c
doc-freshness-reverse-lint is a skill published in the GitHub repository wan-huiyan/claude-ecosystem-hygiene (1 stars, last pushed 25d ago), licensed MIT. It adds 376 tokens to every session and 4,674 once invoked, about $0.0019 per session on Opus 5. A static security scan graded it B with 1 finding (reads agent configuration directories). No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-31.
Other skills, from other repositories
session-handoff
End-of-session handoff that captures session knowledge, dispatches output across the canonical 7-bucket docs/ taxonomy (decisions/runbooks/analysis/references/reviews/handoffs/deliverables — aligned with memory-hygiene v3.3), triggers a doc-freshness reverse-lint + skill-freshness audit to catch stale normative…
llm-wiki
Build and maintain an LLM-curated personal knowledge base — the "LLM Wiki" pattern from Andrej Karpathy's April 2026 gist. Use this skill whenever the user wants to ingest a source (paper, article, transcript, PDF, notes) into a persistent compounding knowledge base, ask a question against accumulated notes, lint or…
reflect
Per-project self-improvement - reads the .harness ledger and feedback memories, then proposes gated rule/threshold/ADR changes so the project stops repeating mistakes. Run periodically.
remember
Review auto-memory entries and propose promotions to CLAUDE.md, CLAUDE.local.md, or shared memory. Also detects outdated, conflicting, and duplicate entries across memory layers.
ucai-patterns
Use when the user asks about Claude Code best practices, how to write agents, how to use hooks, how to manage context, or how to work effectively with Claude Code's native systems.
dream
Memory consolidation - review, merge, prune, and index memory files. Run periodically to keep memories organized and up-to-date.