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 instructions/amal-david/mlx-porting-skill/claude-mdgit clone --depth 1 https://github.com/Amal-David/mlx-porting-skillWrote 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/instructions/amal-david/mlx-porting-skill/claude-md)<a href="https://agentmods.dev/instructions/amal-david/mlx-porting-skill/claude-md"><img src="https://agentmods.dev/badge/instructions/amal-david/mlx-porting-skill/claude-md.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 5.1 | $0.00545 | $0.00545 |
| Opus 5 | $0.00272 | $0.00272 |
| Sonnet 5 | $0.00109 | $0.00109 |
| Haiku 4.5 | $0.00055 | $0.00055 |
Grade C, and why
mlx-porting-skill CLAUDE.md scanned grade C with 1 finding against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured 6d ago.
A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.
Hidden instructionshighPrompt injection
Directives inside HTML comments, invisible characters or bidirectional overrides are read by the model and not by the person reviewing the file.
<!-- superplan-entry-instructions:start --> This is a copy
88% identical to mlx-porting-skill AGENTS.md — 12 lines differ, which has more behind it and is treated as the original. This page carries a canonical link to it rather than competing with it.
How it starts
The opening of the file, as written. The whole thing — 42 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Optional Superplan Operating Contract
When Superplan is installed, load and follow superplan-entry from the first
available location:
.codex/skills/superplan-entry/SKILL.md.superplan/skills/superplan-entry/SKILL.md${HOME}/.config/superplan/skills/superplan-entry/SKILL.md${HOME}/.codex/skills/superplan-entry/SKILL.md
If Superplan is unavailable, continue with the repository-native workflow in
VALIDATION.md; optional orchestration must never make a fresh checkout
unusable.
When Superplan is active:
- No implementation before loading and following
superplan-entry. - No broad repo exploration before loading and following
superplan-entry. - No planning or repo-specific clarification before loading and following
superplan-entry. - Keep workflow control internal: do not narrate skill names, routing, or command logs to the user.
- If
.superplan/exists, treat the Superplan CLI as the execution control plane. - Prefer workspace harnesses, scripts, and custom workflows when
superplan-entryroutes you there.
Canonical loop when Superplan is active:
- Run
superplan status --json. - Claim or resume work with
superplan run --jsonorsuperplan run <task_id> --json. - Continue through the owning Superplan phase instead of improvising a parallel workflow.
- Use lifecycle commands such as
superplan task runtime block,superplan task runtime request-feedback, andsuperplan task review complete; never hand-edit.superplan/runtime/.
Decision guardrails:
- If Superplan readiness is missing, fall back to the repository-native workflow unless the user explicitly asked for Superplan itself.
- If work is already shaped, resume the owning execution or review phase instead of routing from scratch.
- If the request is large, ambiguous, or multi-workstream, route before implementing.
MLX model porting requests
- For any request about porting, converting, running, quantizing, benchmarking, or optimizing a model for MLX or Apple Silicon (e.g. "port this HF model to my Mac", "run Qwen on Apple Silicon", "make this faster on my M3", "fix NaN in my MLX port"), the domain source of truth is
mlx-model-porting/SKILL.md- load it and its Trigger map before improvising. - When available, Superplan governs workflow sequencing; the skill always governs technical content.
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.
- 6d ago First seen · 42 lines · 545 tokens per session scan C 65a50767e91d
mlx-porting-skill CLAUDE.md is an instructions file published in the GitHub repository Amal-David/mlx-porting-skill (5 stars, last pushed yesterday), licensed Apache-2.0. It adds 545 tokens to every session, about $0.0027 per session on Opus 5. A static security scan graded it C with 1 finding (hidden instructions). It is 88% identical to mlx-porting-skill AGENTS.md, differing in 12 lines, and is treated as a copy.
Other instructions, from other repositories
Logue CLAUDE.md
Claude Code instructions for bitwize-ai/Logue, covering logue — development guidelines, project overview, running the build you just made, where documentation goes and dependencies.
Bonsai-demo AGENTS.md
AGENTS.md instructions for PrismML-Eng/Bonsai-demo, covering agent guide — tuning the bonsai demo, why the 27b models (what to show off), the models, knobs that matter (27b) and adding mcp servers.
fitllm-engine AGENTS.md
AGENTS.md instructions for click6067-ship-it/fitllm-engine, covering agents.md — how ai agents should use this repo, use it without cloning, use it as a library, rules for agents modifying this repo and measured data welcome.
tpu_performance_autoresearch_wiki CLAUDE.md
Claude Code instructions for vlasenkoalexey/tpu_performance_autoresearch_wiki, a project described as: Wiki built using Karpathy method containing information about TPU performance optimizations and hooking it up to autoresearch optimization engine.
tuning-engines-cli CLAUDE.md
Instructions for cerebrixos-org/tuning-engines-cli, covering tuning engines mcp server, setup, claude desktop configuration, claude code and available tools.
vllm.cpp CLAUDE.md
Claude Code instructions for mudler/vllm.cpp, a project described as: a community oriented 1:1, vLLM-alike (Continuous batching, paged KV) engine in C++ with additional features (GGUF, RadixAttention, Cache-aware scheduling, ...).