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 Mark-Life/agent-skills --skill human-to-agentgit clone --depth 1 https://github.com/Mark-Life/agent-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/mark-life/agent-skills/human-to-agent)<a href="https://agentmods.dev/skills/mark-life/agent-skills/human-to-agent"><img src="https://agentmods.dev/badge/skills/mark-life/agent-skills/human-to-agent/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/mark-life/agent-skills/human-to-agent"><img src="https://agentmods.dev/badge/skills/mark-life/agent-skills/human-to-agent.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 | $0.00054 | $0.00680 |
| Opus 5 | $0.00027 | $0.00340 |
| Sonnet 5 | $0.00011 | $0.00136 |
| Haiku 4.5 | $0.00005 | $0.00068 |
Grade A, and why
human-to-agent 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 10d 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 — 65 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Human → Agent
You are writing for a stochastic reader. Every instruction is a lever on how predictably it behaves. The aim is predictability — the agent taking the same process each run, not producing the same output. Each rule below buys some.
Prompt the positive
State the behaviour you want. Prohibition backfires: don't think of an elephant names the elephant and makes it more available, not less. Phrase the target so the banned behaviour is never spoken. Keep a "don't" only as a hard guardrail you cannot phrase positively, and even then pair it with what to do instead.
Leading words
A leading word is a compact concept the model already holds from pretraining — tracer bullet, fog of war, tight loop, receipts. One well-chosen word anchors a whole region of behaviour in a token, because it recruits priors the model already has. Repeat it and its meaning accretes; use the same word across your prompts, docs, and code so the agent links them and fires the right behaviour.
Reach for one wherever you have spelled a quality out three times: "fast, deterministic, low-overhead" is one idea restated — collapse it to tight.
Cut no-ops
A no-op is a line the model already obeys by default: it costs context and buys nothing. Test each sentence in isolation — does it change behaviour versus the default? "Be thorough", when the model is already thorough-ish, is a no-op; the fix is a stronger word (relentless), not more words. When a sentence fails, delete the whole sentence rather than trim it.
One source of truth
Keep each instruction in exactly one place. The same rule in two files (duplication) rots the moment one copy changes, and inflates the rule's apparent weight. When you change the behaviour, you want a one-place edit.
Say how "done" is checked
Give the agent a completion criterion it can check — done from not-done, observable. Where it matters, make it exhaustive: "every changed file accounted for", not "list the changes". A vague criterion invites the agent to stop early.
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.
- 10d ago First seen · 65 lines · 54 tokens per session scan A 72d0fe0bb03d
human-to-agent is a skill published in the GitHub repository Mark-Life/agent-skills (2 stars, last pushed 25d ago), licensed MIT. It adds 54 tokens to every session and 680 once invoked, about $0.0003 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
ralplan
Consensus planning entrypoint that auto-gates vague ralph/autopilot/team requests before execution.
remember
Review reusable project knowledge and decide what belongs in project memory, notepad, or durable docs.
embedding-strategies
Select and optimize embedding models for semantic search and RAG applications. Use when choosing embedding models, implementing chunking strategies, or optimizing embedding quality for specific domains.
docx-comment-reply
Reply to comments (批注) in Word .docx/.doc files: extract comment context, draft replies, write threaded replies back, and validate OOXML.
explaining-machine-learning-models
Explain trained machine learning models through feature attribution, local explanations, and behavior summaries. Use as an explicit/manual helper once a model already exists, not for training ownership, leakage auditing, or general ML strategy selection.
config-validator
Validate AIWG configuration files and project setup for correctness and completeness.