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 skills add CaesiumY/dding-dong --skill claude-md-optimizergit clone --depth 1 https://github.com/CaesiumY/dding-dongWrote 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/caesiumy/dding-dong/claude-md-optimizer)<a href="https://agentmods.dev/skills/caesiumy/dding-dong/claude-md-optimizer"><img src="https://agentmods.dev/badge/skills/caesiumy/dding-dong/claude-md-optimizer/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/caesiumy/dding-dong/claude-md-optimizer"><img src="https://agentmods.dev/badge/skills/caesiumy/dding-dong/claude-md-optimizer.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 | $0.00118 | $0.01397 |
| Opus 5 | $0.00059 | $0.00698 |
| Sonnet 5 | $0.00024 | $0.00279 |
| Haiku 4.5 | $0.00012 | $0.00140 |
Grade A, and why
claude-md-optimizer 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 9d 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 — 145 lines — stays where its author put it; the contents beside it link to each section on GitHub.
CLAUDE.md Optimizer
Optimize CLAUDE.md files by applying the discoverability filter: remove information agents can discover from code, keep only non-discoverable operational knowledge (gotchas, landmines, non-standard conventions), and mine source code for undocumented gotchas.
Research shows redundant context (directory trees, data flow diagrams) degrades agent performance by 15-20%, while human-authored operational knowledge reduces runtime by ~28%.
Flag Parsing
Parse $ARGUMENTS for optional flags:
| Flag | Effect |
|---|---|
--dry-run |
Analyze and show diff without modifying the file |
--report-only |
Output statistics and classification table only |
--path <path> |
Target CLAUDE.md path (default: ./CLAUDE.md) |
--help |
Display usage and exit |
If --help is present, display available flags and a brief description of the 4-step workflow, then stop.
Workflow
Step 1: Baseline Analysis
Read the target CLAUDE.md. Collect line statistics:
node .claude/skills/claude-md-optimizer/scripts/line-count.mjs '<TARGET_PATH>'
Classify each ##/### section into one of three categories:
| Category | Meaning | Action |
|---|---|---|
discoverable |
Agent can find this via Glob/Grep/Read within 10 seconds | Remove |
operational |
Non-discoverable, operationally significant | Keep |
verbose |
Operational knowledge but overly detailed | Compress |
To classify, actually read the source files referenced in each section. Verify whether the information is truly discoverable. Detailed classification criteria are in references/methodology.md.
Present results as a table:
## Baseline Analysis — <filename>
Total: XXX lines (YY sections)
| Section | Lines | Category | Rationale |
|---------|-------|----------|-----------|
| Directory Structure | 63 | discoverable | Glob **/* reveals this instantly |
| Design Rules | 8 | operational | Non-standard constraints, not in code |
| Config & State | 24 | verbose | Operational but compressible to ~6 lines |
Removal candidates: XX lines (XX%)
What ships with it
4 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.
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.
- 9d ago First seen · 145 lines · 118 tokens per session scan A bca8e1faa993
claude-md-optimizer is a skill published in the GitHub repository CaesiumY/dding-dong (2 stars, last pushed 6mo ago), licensed MIT. It adds 118 tokens to every session and 1,397 once invoked, about $0.0006 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 skills, from other repositories
engram
Read and write the project's durable reasoning memory (decisions, principles, cautions, problems, insights) through the Engram MCP tools. Recall relevant memory before non-trivial work; capture only durable, high-value knowledge silently at natural stopping points; keep the graph honest (judge suspects, close answered…
engram-digest
Seed or top up this project's Engram graph from the existing codebase — an explicit, user-invoked digestion of the current working tree into typed memory nodes (decisions, principles, cautions, problems, insights, intents). Use when the user says "digest this project", "ingest the codebase into memory", "seed the…
persona-triggers
A detailed guide for matching Korean and English trigger phrases to working viewpoints, such as security, testing, design, operations, and AI. A trigger phrase is wording in a request that activates a particular response approach.
persona-format
A Korean-language response format for coding, debugging, and implementation work. It asks for structured reasoning, multiple expert viewpoints, a recommended solution, execution steps, verification, and prevention of repeat problems.
persona-marketer
A review guide for marketing and sales work, including copywriting, content, search visibility, advertising, landing pages, and customer retention. It looks at the work from the viewpoint of an experienced marketer and salesperson.
persona-safety
A safety guide for coding, user-interface, security, deployment, deletion, and force-push work. It requires extra checks for secrets, sensitive data, risky changes, cleanup paths, and mobile layouts.