Getting it into your agent
It runs from inside its repository, so the clone comes first — what it calls does not travel with the file alone.
git clone --depth 1 https://github.com/allemaar/open-skillsnpx agentmods add skills/allemaar/open-skills/new-skill-creatorWrote 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/allemaar/open-skills/new-skill-creator)<a href="https://agentmods.dev/skills/allemaar/open-skills/new-skill-creator"><img src="https://agentmods.dev/badge/skills/allemaar/open-skills/new-skill-creator/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/allemaar/open-skills/new-skill-creator"><img src="https://agentmods.dev/badge/skills/allemaar/open-skills/new-skill-creator.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 | $0.00182 | $0.05952 |
| Opus 5 | $0.00091 | $0.02976 |
| Sonnet 5 | $0.00036 | $0.01190 |
| Haiku 4.5 | $0.00018 | $0.00595 |
Grade B, and why
new-skill-creator scanned grade B with 1 finding against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured 11d ago.
A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.
Enumerates other installed skillsmediumAgent snooping
Other skills' SKILL.md files reveal prompts, capabilities and secrets that should be invisible to peers.
2. Snapshot symlink state — POSIX: `ls -la ~/.claude/skills ~/.codex/skills ~/.agents/skills > "$TMPDIR/symlinks-before.txt"`. Windows PowerShell: `Get-ChildItem ~\.claude\skills, ~\.codex\skills, ~\.agents\skills | Out- How it starts
The opening of the file, as written. The whole thing — 297 lines — stays where its author put it; the contents beside it link to each section on GitHub.
/new-skill-creator
What this skill does
This is a maintainer's authoring loop, and it assumes a skills repo you own. It writes a new skill into your clone, links that clone into your runtimes, and pushes. The linking below is the opposite of what
README.mdtells you to do when installing this pack — and deliberately so. A symlink means your agent runs whatever the nextgit pullbrings, with no moment where you read the diff; that is the driftTHREAT-MODEL.mdnames, and it is why copying is the default there. Linking here does not make that drift go away — it accepts it, on the same terms README already states: you're maintaining the pack, or you read every pull. Authoring a skill does not pre-review a contributor's later commit to it, or your own work from another machine. Installing skills someone else maintains is a different act — copy those.
Creates a new skill in your skills repo, links it into all three agent runtimes (Claude Code, Codex, and the shared .agents/ dir), commits and pushes to GitHub. Supports two formats:
- Single-doc — one
SKILL.md(default for short routers, triggers, mode-setters, reference cards). - Dual-doc —
SKILL.md(human-readable, self-sufficient) +protocol.yon(declarative, mechanically validatable protocol) (default for heavy procedural skills with phases, gates, severity-ranked rules). See section 2b.
Structured execution spec:
protocol.yon. Read it for the canonical rules and step sequence; this file is explanation. The two must stay in sync — if you edit one, update the other and refresh the@STAMPdate.
Steps
1. Gather info
Ask the user (on Claude Code, use AskUserQuestion for structured options; on Codex or other agents without that tool, ask via prose with options as a numbered list):
- What should this skill be called? (kebab-case, becomes /skill-name)
- What does it do — one sentence?
- When should it trigger? Give 2-3 example phrases a user might say.
- Could this skill be confused with an existing skill? If yes, list the ones it's most similar to — we'll add boundary clauses to clarify.
- Should only the user invoke it (not Claude auto-trigger)? Default: yes for workflows with side effects.
- Is this skill runtime-specific? Default: no — skills are available in all three targets (Claude Code, Codex, and the shared
.agents/dir). Only mark runtime-specific if the skill depends on tools or capabilities that are not available in the other runtime(s). - Dual-doc or single-doc? Apply the criterion below; default to single-doc if uncertain.
- Caller Options participant? If the skill has genuine venue or mode optionality worth surfacing to the caller, add the COP opt-in block — see
caller-options/references/OPT-IN-BLOCK.md. - Next-skills participant? If, after this skill finishes, there are natural successor skills worth recommending to the user, add the NSP opt-in block — see
next-skills/SKILL.mdand the subsection below. - Self-improvement participant? If this is a capable procedural skill (phases, gates, judgment) that a run could expose a tunable weakness in, add the SIP opt-in block (
self-improvable: true) — seeself-improve/SKILL.mdand the subsection below. Routers / mode-setters / reference cards omit it. - Public or private? Decide
visibility:—privateif the skill couples to personal infrastructure or proprietary IP (vault paths, personal domains, proprietary engines);publicif it is generic and reusable.
What ships with it
3 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.
- 11d ago First seen · 297 lines · 182 tokens per session scan B 0cb0816cdd22
new-skill-creator is a skill published in the GitHub repository allemaar/open-skills (14 stars, last pushed yesterday), licensed Apache-2.0. It adds 182 tokens to every session and 5,952 once invoked, about $0.0009 per session on Opus 5. A static security scan graded it B with 1 finding (enumerates other installed skills). No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-30.
Other skills, from other repositories
polis-protocol
Set up and run a self-improving, multi-vendor AI agent team with the Polis Protocol — a markdown polis/ folder where each agent is a citizen with a capability card, tasks are contracts routed to whoever has the best track record by a learning router, settled work files lessons that compound into team memory, and…
media
Skill "media" from ymxlx/polis-protocol, covering polis protocol: a self-optimizing city of agents, the core idea, when this skill is active, structure of a polis and the first thing to do every session.
_data
Skill "_data" from ymxlx/polis-protocol, covering polis protocol: a self-optimizing city of agents, the core idea, when this skill is active, structure of a polis and the first thing to do every session.
subagent-driven-development
Execute plans via delegatetask subagents (2-stage review).
mcporter
List, auth, and call MCP servers/tools from the terminal.
research-engineer
An uncompromising Academic Research Engineer. Operates with absolute scientific rigor, objective criticism, and zero flair. Focuses on theoretical correctness, formal verification, and optimal implementation across any required technology.