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 mh-language/mhl-core-runtime --skill mhl-toolsgit clone --depth 1 https://github.com/mh-language/mhl-core-runtimeWrote 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/mh-language/mhl-core-runtime/mhl-tools)<a href="https://agentmods.dev/skills/mh-language/mhl-core-runtime/mhl-tools"><img src="https://agentmods.dev/badge/skills/mh-language/mhl-core-runtime/mhl-tools/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/mh-language/mhl-core-runtime/mhl-tools"><img src="https://agentmods.dev/badge/skills/mh-language/mhl-core-runtime/mhl-tools.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 | $0.00051 | $0.02020 |
| Opus 5.5 | $0.00020 | $0.00808 |
| Sonnet 5.5 | $0.00010 | $0.00404 |
| Haiku 4.5 | $0.00005 | $0.00202 |
Grade A, and why
mhl-tools 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 today.
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 — 224 lines — stays where its author put it; the contents beside it link to each section on GitHub.
mhl tools and integrations
Assumes the mhl-language skill. Exact signatures are in the mhl-language skill's
references/stdlib.md.
The failure convention
Native ops fail in two different ways. Choose your handling per op:
| Op | On failure | You must |
|---|---|---|
cmd.exec, git.status, git.add, git.commit |
returns {stdout, stderr, exit_code} |
check exit_code |
http.* non-2xx |
returns {status, ok, body, ...} |
check .ok, or set raise_for_status: true |
http.* transport error, fs.read, json.parse, git.diff, git.log, git.rev_parse |
raises | try only if you can recover |
An unchecked cmd.exec is the classic silent bug: the step passes while the command
failed. Wrap it once and reuse the wrapper.
Declaring tools
A tool is a namespace of methods that gives steps, hooks and tests one reusable,
testable API over native ops. Put logic here and keep steps thin.
tool Shell {
// Fails the step on a non-zero exit, so callers need no check.
must(argv: string[]): string {
var result = cmd.exec(argv, timeout: 30s)
if (result.exit_code != 0) fail(self.describe_failure(argv, result.stderr))
return result.stdout.trim()
}
// internal: callable through self. and from this file's tests only.
internal describe_failure(argv: string[], stderr: string): string ->
"${argv.join(" ")} failed: ${stderr.trim()}"
}
test shell {
describe returns_trimmed_stdout {
are_equal(Shell.must(["echo", "hi"]), "hi")
}
describe nonzero_exit_fails {
var message = ""
try {
Shell.must(["sh", "-c", "echo nope >&2; exit 3"])
} catch (e) {
message = "${e}"
}
is_true(message.contains("nope"))
}
describe internal_helpers_are_testable_here {
are_equal(Shell.describe_failure(["ls"], "x\n"), "ls failed: x")
}
}
- Annotate parameters and returns. Calls are checked by lint where it can see the types.
- Bind arguments by position, in declaration order. Names in a tool call are ignored.
- Use an expression body (
name(x): T -> expr) or a block (name(x): T { ... }). - Keep control-flow exits (
goto,pause,complete,break) in steps.fail()belongs in tools freely. - Share a value across methods with
constorvarin the tool body, and compose withself.other(...). Write durations as literals at the use site.
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.
- today First seen · 224 lines · 51 tokens per session scan A bd343dd663d9
mhl-tools is a skill published in the GitHub repository mh-language/mhl-core-runtime (9 stars, last pushed today), licensed MIT. It adds 51 tokens to every session and 2,020 once invoked, about $0.0002 per session on Opus 5.5. 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 2026-11-07.
mhl-tools compared with skills from other repositories
| Mod | Type | Stars | Last push | Licence | Scan | Tokens | Works with |
|---|---|---|---|---|---|---|---|
| mhl-toolsmh-language/mhl-core-runtime | Skill | 9 | today | MIT | A | 51 | Claude Code, Codex |
| git-ai-searchgit-ai-project/git-ai | Skill | 2.9k | 2d ago | Apache-2.0 | A | 13 | Claude Code |
| rememberpacifio/atlas | Skill | 9.5k | 2d ago | Apache-2.0 | A | 32 | Claude Code |
| releasematrixorigin/memoria | Skill | 623 | 2d ago | Apache-2.0 | A | 27 | Claude Code, Codex |
| recall-rationaletobihagemann/turbo | Skill | 419 | 14d ago | MIT | A | 74 | Claude Code, Codex |
| change-boottikalk/adlc-team-skills | Skill | 144 | yesterday | MIT | A | 113 | Claude Code, Codex |
| jolli-searchjolliai/jolliai | Skill | 149 | 3d ago | Apache-2.0 | A | 39 | Claude Code, Codex |