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 brody-0125/my-claude-skills --skill meta-rulesgit clone --depth 1 https://github.com/brody-0125/my-claude-skillsWrote 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/brody-0125/my-claude-skills/meta-rules)<a href="https://agentmods.dev/skills/brody-0125/my-claude-skills/meta-rules"><img src="https://agentmods.dev/badge/skills/brody-0125/my-claude-skills/meta-rules/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/brody-0125/my-claude-skills/meta-rules"><img src="https://agentmods.dev/badge/skills/brody-0125/my-claude-skills/meta-rules.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 | $0.00033 | $0.00876 |
| Opus 5 | $0.00016 | $0.00438 |
| Sonnet 5 | $0.00007 | $0.00175 |
| Haiku 4.5 | $0.00003 | $0.00088 |
Grade A, and why
meta-rules 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 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.
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 — 83 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Meta-Rules — Plugin Component Quality Gates
Rules that govern how agents, skills, and workflows should be written. All auto-generated improvements MUST pass these rules before application.
Token Limits
These limits apply to auto-generated improvements (output of improvement-generator and auto-optimizer). They constrain individual components, not the main orchestrating SKILL.md which serves as a command router.
| Component | Max Tokens | Measurement | Scope |
|---|---|---|---|
| Agent description (.md) | 1000 | chars/4 | Per agent file |
| Knowledge skill (SKILL.md) | 500 | chars/4 | Per knowledge skill (meta-rules, analysis-patterns, cost-tracking) |
| Main skill (SKILL.md) | No hard limit | — | Orchestrating skill — uses lazy-load via resources/ |
| Single example | 200 | chars/4 | Per example block |
| Hook script | 100 lines | line count | Per standalone script |
Writing Rules
Agents
- ALWAYS use imperative voice in instructions ("Analyze the traces" not "You should analyze the traces")
- ALWAYS specify input and output formats explicitly
- NEVER exceed 1000 tokens in the agent description
- ALWAYS include model assignment (haiku/sonnet/opus) with justification
- Include 1-2 examples maximum — more examples waste context
- NEVER use vague instructions ("do your best", "try to", "if possible")
- ALWAYS specify exit conditions — when is the agent done?
Skills
- ALWAYS keep SKILL.md under 500 tokens — use resources/ for details
- ALWAYS use the 2-layer structure: SKILL.md (core) + resources/ (on-demand)
- NEVER load all resources at once — lazy-load per phase/command
- ALWAYS declare allowed-tools explicitly — minimal permissions
- ALWAYS include trigger conditions in the description
Hooks
- ALWAYS complete within 50ms — hooks block execution
- ALWAYS use
|| truefallback — hook failures must not break workflow - NEVER read large files in hooks — use append-only writes
- ALWAYS use
set -euo pipefailin scripts - NEVER install external dependencies — bash + jq only
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 · 83 lines · 33 tokens per session scan A 22ffa2a3917d
meta-rules is a skill published in the GitHub repository brody-0125/my-claude-skills (2 stars, last pushed 6mo ago), licensed MIT. It adds 33 tokens to every session and 876 once invoked, about $0.0002 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
self-audit
Self-audit this skills repo against CLAUDE.md invariants, the artifact contract, and README/CLAUDE.md/docs sync via three parallel read-only subagents. Local meta-skill — independent of the /task: pipeline.
grill
Interrogate a plan or decision one question at a time before capture, keeping a decision-plus-rationale ledger, then route to the right capture skill.
architect
Answer build-or-adopt honestly, then design one committed architecture from your own understanding of the problem — driven by ranked quality attributes, stress-tested against the field and against change, proven by a walking skeleton, its consequences confirmed by the user, committed to ARCHITECTURE.md. A command the…
designer
Settle the look before the pixels — audience and context first, then the base and the delta, taste settled by looking at real states, accessibility as the floor, one design language committed to DESIGN.md as tokens with roles. A command the user types, for work with an interface someone will see; standalone, no other…
daily-working
End-to-end pipeline: pull a task from the project's task tracker by ID (Redmine or GitHub Issues today, more addable via a new adapter), sanity-check and impact-assess it against the codebase before touching anything, implement it with the Claude CLI, verify the result in a real browser via the Claude Chrome extension…
compact
Put the compaction question to a snapshot that drifted — route what no longer constrains unfinished work, and say what moves before moving it. In flight or done; the log is appended to, never edited or trimmed.